Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 1106

_____________________________ _____________________________

Tumor Stroma in Anaplastic Thyroid Carcinoma

Interstitial Collagen and Tumor Interstitial Fluid Pressure

BY

ELLEN LAMMERTS

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2001

Dissertation for the Degree of Doctor of Philosophy (Faculty of Medicine) in MedicalBiochemistry presented at Uppsala University in 2002

AbstractLammerts, E. 2001. Tumor stroma in anaplastic thyroid carcinoma. Interstitial collagenproduction and tumor interstitial fluid pressure. Acta Universitatis Upsaliensis.Comprehensive Summaries of Uppsala Dissertations from the Faculty of Medicine1106. 50 pp. Uppsala. ISBN 91-554-5198-5

Anaplastic thyroid carcinoma (ATC) is an aggressive malignancy in man with stromalfibrosis as one of the main features. Carcinoma cells synthesized no or little collagen Iprotein. Pro-α1(I) collagen mRNA was expressed by stromal cells throughout thetumor, but expression of procollagen type I protein was restricted to stromal cellssituated close to nests of carcinoma cells. These data suggest that the carcinoma cellsstimulated collagen type I deposition by increasing pro-α1(I) collagen mRNAtranslation.

Cocultures, of the human ATC cell line KAT-4, with fibroblasts underconditions that allow the study of stimulatory factors on collagen mRNA translation,showed that the KAT-4 cells stimulated collagen type I protein synthesis in fibroblasts.Specific inhibitors of PDGF and TGF-β1 and -β3 were able to inhibit this carcinomacell-induced stimulation of collagen type I synthesis. These findings suggest that tumorcells were able to stimulate collagen mRNA translation in stromal fibroblasts by, atleast in part, transferring PDGF and/or TGF-β1 and –β3.

Xenograft transplantation of different ATC cell lines into athymic micedemonstrated that the low collagen producing carcinoma cell lines were lesstumorigenic compared to non-collagen producing carcinoma cell lines. The morphologyof tumors derived from non-collagen producing ATC cell lines showed a well demarkedstroma surrounding carcinoma cell nests.

TGF-β1 and –β3 were found to play a role in generating a high tumor interstitialfluid pressure (TIPF) in experimental KAT-4 tumors. A specific inhibitor of TGF-β1and –β3 was able to lower TIPF and reduce tumor growth after a prolonged period oftreatment, suggesting that TGF-β1 and –β3 have a role in maintaining a stroma thatsupport tumor growth.

Key words: Collagen I synthesis, Fibrosis, Tumor – stroma interactions, PDGF, TGF-β,Inhibitor, Cell cycle, Tumor growth, Hyaluronan.

Ellen Lammerts, Department of Medical Biochemistry and Microbiology, BiomedicalCenter, Box 582, SE-751 23 Uppsala Sweden

Ellen Lammerts 2001

ISSN 0282-7476

ISBN 91-554-5198-5

Printed in Sweden by Lindbergs Grafiska HB, Uppsala 2001

Papers

This thesis is based on the following papers, which will be referred to in the text bytheir Roman numerals:

I. T. Dahlman, E. Lammerts, M. Wik, J. D. Bergström, L. Grimelius, K.Westermark, K. Rubin, N.-E. HeldinFibrosis in undifferentiated (anaplastic) thyroid carcinomas: evidence for a dualaction of tumour cells in collagen type I synthesisJournal of Pathology, 191: 376-386, 2000

II. T. Dahlman, E. Lammerts, J. D. Bergström, Å. Franzén, K. Westermark, N.-E.Heldin, K. RubinCollagen type I expression in experimental anaplastic thyroid carcinoma:regulation and relevance for tumorigenicityIntternational Journal of Cancer (in press), 2001

III. E. Lammerts, P. Roswall, C. Sundberg, P. J. Gotwals, V. E. Koteliansky, N.-E. Heldin, K. Rubin

A soluble inhibitor of TGF-β1 and -β3 reduces tumor interstitial fluid pressureand promotes tumor growth in experimental anaplastic thyroid carcinomaManuscript, 2001

IV. E. Lammerts, A. Jacobson, A. Salnikov, P. Roswall, C. Sundberg, N. Khorasani,P. Heldin, N.-E. Heldin, K. Rubin

A soluble TGF-β1 and -β3 inhibitor lowers collagen type I deposition inexperimental anaplastic thyroid carcinomaManuscript, 2001

Reprints were made with permission from the publisher.

Contents

Abbreviations 6

Introduction 7

Functions of the tumor stroma 8Tumor stroma generation 9Wound healing, a model for tumor stroma generation 10Tumor stroma composition 11Collagen type I 13Thyroid cancer 14

TGF-β and TGF-β signaling 17

TGF-β inhibition of the cell cycle 20PDGF and PDGF signaling 21

The biological activities of PDGF and TGF-β 23Interstitial fluid pressure 25Tumor interstitial fluid pressure 26

Aims 28

Methodology 29

ATC cell lines 29Quantification of collagen synthesis 29Cell cultures on poly[HEMA] 29Immunohistochemistry and in situ hybridization 30Measurement of TIFP 31

Results 32

Paper I 32Paper II 33Paper III 33Paper IV 34

Conclusions and future perspectives 35

Conclusions 35Future perspectives 35

Acknowledgements 36

References 38

Abbreviations

Anaplastic thyroid carcinoma ATCbasic Fibroblast Growth Factor bFGFCapillary hydrostatic pressure CHPColloid osmotic pressure in the interstitial fluid COPIF

Colloid osmotic pressure in plasma COPPL

Epidermal Growth Factor EGFEpithelial to mesenchymal transition EMTExtracellular matrix ECMGlycosaminoglycans GAGsHepatocyte Growth Factor HGFHyaluronan HAHydroxyproline HPInterstitial fluid pressure IFPPlasminogen activator inhibitor PAIPlatelet-Derived Growth Factor PDGFTissue inhibitor of metallo proteinases TIMP

Transforming Growth Factor-β TGF-βTumor interstitial fluid pressure TIFP

Tumor Necrosis Factor-α TNF-αVascular Endothelial Growth Factor VEGFVascular Permeability Factor VPF

7

Introduction

Cancer is still a major cause of death in the developed countries. Knowledge of thegenetical changes that lead to the development of a malignant cell has increaseddramatically during the last decades. This progress has, however not been paralleled bynovel and likewise dramatically improved treatment regimes. Most cancers are stilltreated by surgical resection, radiation or chemotherapy and combinations thereof.During the last 10 years it is has been increasingly recognized that tumor stroma playsan important part in the determination of the tumor phenotype and that the stromacompartment is essential for the outgrowth of a tumor beyond 1 mm3 (1-3). Thisknowledge has initiated new fields of research that might improve cancer treatment. Thetumor stroma provides the tumor with structural organization, blood vessels andinterstitial connective tissue. Inhibition of angiogenesis is an attractive therapeutictarget. Modulating the hydraulic properties of the tumor stroma to improve uptake ofanti-cancer drugs, constitutes another target.

During the development of malignant cancers, cells escape the normal regulatoryprocesses and gain unlimited potential growth and the ability to invade surroundingtissues. Mutations are key events in the initiation of cancer development. Most tumorsdisplay one or more mutations/gene aberrations resulting in loss of tumor suppressorgene function, control of transmembrane receptor function, and/or signaling pathways.The malignant cells in a rapidly growing tumor have acquired 6 characteristics: 1) self-sufficiency in growth signals; 2) insensitivity to anti-growth signals; 3) ability to escapeprogrammed cell death; 4) ability to show unlimited proliferation; 5) ability to inducesustained angiogenesis and 6) ability to invade tissues and metastasize (4). The numberof reports about different mutations in tumors are numerous and beyond the scope ofthis thesis (see also reviews: (5, 6)).

Cancers are classified according to the tissue and cell type of origin. Cancers fromepithelial cells are termed carcinomas and mesenchymal malignancies are termedsarcomas. Other cancer types do not fit in these two categories, such as leukemia,melanoma and cancers derived from cells of the nervous system. The vast majority ofhuman cancers are carcinomas. Malignant carcinoma cells have the ability to breakloose, invade surrounding tissues, enter the blood stream or lymphatics and formsecondary tumors at distant sites in the body. The ability to metastasize is the one singlefactor that has the highest impact on prognosis. Different types of carcinoma vary indegree of malignancy. Basal cell carcinoma, for example, is only locally invasive, rarelyforms metastases, and has a good prognosis if treated, whereas anaplastic thyroid

8

carcinoma (ATC) is much more malignant, rapidly gives rise to many metastases and isoften impossible to treat (7-9).

A carcinoma represents a complex ecosystem where multiple cell – extracellularmatrix (ECM) and carcinoma cells – stroma interactions provide reciprocal influences,which result in carcinoma cell promotion, invasion and metastasis.

Functions of the tumor stroma

Solid tumors are composed of two distinct but interdependent compartments: thecarcinoma cells themselves and the supporting stroma that they induce and in whichthey are dispersed. Tumor stroma differs from the normal interstitial loose connectivetissue in many important aspects and is believed to be involved in malignant growth.Components of the ECM provide a variety of different signals, which influence cellgrowth, migration, invasion, differentiation and biosynthetic activities (10, 11). Inaddition, the ECM plays important roles in cell survival, and loss of contacts with ECMcomponents results in apoptosis in normal cells (12).

Tumor angiogenesis occurs in several steps. Degradation of pre-existing vesselbasement membranes, proliferation and directed migration of endothelial cells towardsan angiogenic stimulus are followed by differentiation of new capillary loops andestablishment of a functional circulation (13). Vascular endothelial growthfactor/vascular permeability factor (VEGF/VPF) and fibroblast growth factor (FGF) arethe main inducers of tumor vascularization and are actively secreted by a variety ofhuman tumor cells (14, 15).

During invasion stroma is involved in both matrix degradation and tumor cellmigration along different components of the ECM. Some fibroblasts of the tumorstroma are characterized by a myofibroblastic phenotype, based on their expression of

α-smooth muscle actin (16, 17). These cells are also capable of producing proteolyticenzymes, such as matrix metalloproteineases (MMPs) that are involved in thedegradation of ECM components, which is one of the prerequisites for invasion. Theseenzymes can be secreted by carcinoma cells however, it has become clear that a majorfraction is produced by stromal cells (18), . Once synthesized, matrix degradingproteases are secreted as inactive pro-enzymes which must be locally activated by theUrokinase type plasminogen activator (uPA)/plasmin system of serine proteases (19).The activity of the stromal proteases is tightly controlled via the synthesis of differentinhibitors of serine proteases, such as plasminogen activator inhibitor (PAI 1 and 2) and

9

tissue inhibitors of metalloproteinases, (TIMP 1 and 2). Such inhibitors can be producedby stromal cells and stored in the ECM (20-24, 25 ).

Tumor stroma generation

Fibrin deposition was one of the earliest morphological events after tumortransplantation in normal tissue (reviewed in: (26)). Fibrin appears to serve as aprovisional stroma that is gradually replaced by granulation tissue and then by maturestroma. Most tumors are supplied by a microvasculature lined by a continuous non-fenestrated endothelium that would not permit the passage of significant amounts ofplasma proteins. Nevertheless, host vessels and newly formed vessels at the tumor siteare permeable for biologically active plasma proteins, including fibrinogen, fibronectin,plasminogen and various clotting factors. Plasma leakage probably occurs throughtrans-endothelial pathways and not via inter-endothelial gaps (27-30).

Once plasma extravasates from the blood vessel the extrinsic or tissue factorclotting pathway is activated, depositing cross-linked fibrin (27, 31). A permeabilityfactor (VPF/VEGF), seems to be responsible for the tumor vessel hyperpermeability(14). VPF/VEGF is a peptide of 34-43 kDa and is 50.000 times more active thanhistamine in enhancing microvascular permeability. Coagulation is mediated by aninteraction between extravasated plasma clotting factors and tumor-associated andperhaps other tissue pro-coagulants. The fibrin turnover in tumors is extensive due tothe presence of plasmin, the principal fibrin-degrading proteinase (32, 33). Fibrinnonetheless accumulates in amounts variable from tumor to tumor, that are sufficient toprovide a provisional stroma. The cross-linked fibrin provides a substrate for theimmigration of new vascular endothelium, fibroblasts, macrophages and otherconnective tissue cells. These events are partly regulated by cytokines and growthfactors produced by inflammatory cells as well as by tumor cells (34). As mentionedbefore tumor stroma is different from normal interstitial connective tissue. This isreflected in a different composition of ECM proteins, which will be further described insection tumor stroma components.

10

Wound healing; a model for tumor stroma generation

Wound healing has been the subject of extensive studies and has several similarities totumor stroma formation. Wound healing encompasses several phases that aretemporarily overlapping, like inflammation, tissue formation and tissue remodeling.

Tissue injury results in extravasation of plasma and of cellular components from thedamaged blood vessels. The blood clot re-establishes hemostasis and provides aprovisional ECM for cell migration. Platelets contribute soluble clotting factors but alsosecrete several mediators of wound healing, such as platelet-derived growth factor(PDGF), that attract and activate macrophages and fibroblasts (35). Numerousvasoactive mediators and chemotactic factors are generated by coagulation and theyrecruit inflammatory leucocytes to the wound (36). Infiltrating neutrophils clean thewound area of foreign particles and bacteria followed by the infiltration and activationof monocytes into macrophages that start to secrete growth factors, such as PDGF andVEGF, which initiate the formation of granulation tissue. Activated macrophages playan important role in this phase (37), by remodeling the provisional matrix. Adherence tothe ECM also activates monocytes to change into inflammatory and reparativemacrophages. The cytokines secreted by monocytes and macrophages play an importantrole in the initiation and propagation of new tissue formation in wounds (38-41).

New stroma or granulation tissue begins to invade the wound area approximatelyfour days after injury. Macrophages, fibroblasts and blood vessels move into the woundspace at the same time (42). Cell movement into a blood clot of cross-linked fibrin mayrequire an active proteolytic system that can cleave a path for cell migration. A varietyof fibroblast derived enzymes, such as plasminogen activator, collagenases, gelatinaseA (MMP2), and stromelysin (MMP3) are involved (43, 44). The macrophages provide aongoing source of growth factors necessary to stimulate fibroplasia and angiogenesis(41, 45); the fibroblasts produce the new ECM; and blood vessels carry oxygen andnutrients necessary to sustain cell metabolism. The structural molecules of newlyformed ECM contribute to the formation of granulation tissue by providing a scaffoldfor cell migration. These molecules include fibrin, fibronectin and hyaluronic acid (46,47). The granulation tissue is gradually replaced with a collagenous matrix (36, 48).Once collagen has been deposited in the wound, the fibroblasts stop producing collagen,and a relatively acellular scar replaces the fibroblast-rich granulation tissue.

Despite the similarities between tumor stroma generation and wound healing thereare also important differences. In wounds, platelets have a central role, releasingclotting factors and other types of mediators, like PDGF, that excerts multiple effects oninflammatory and connective tissue cells (49-51). No such role for platelets has been

11

demonstrated in tumor stroma generation. However carcinoma cells themselves expressclotting factors and secrete growth factors (51-53). A second major difference is that thevascular hyperpermeability is perpetual in tumors and self-limited in wounds. Woundsusually heal within a defined period of time. Healing is closely linked to the cessation oflocal vascular hyperpermeability (54). Some parts of the tumor appear to undergo ahealing-like process, sometimes even showing a scar-like desmoplasia. However,invading carcinoma cells keep the ability to render local blood vessels hyperpermeableand stimulate adjacent connective tissue cells, continuing the process of extravasationand tumor stroma growth (31).

Tumor stroma composition

Tumor stroma is composed of inflammatory and connective tissue components that arederived from the circulating blood, from the adjacent host tissue and from newly formedblood vessels (53, 55, 56). Blood derived components include plasma and various typesand numbers of inflammatory cells. Many of the elements found in normal connectivetissue are also present in tumor stroma (57).

The tumor stroma extracellular matrixThe cells present in the tumor stroma are responsible for the production of ECMmacromolecules. In loose interstitial connective tissues the ECM components areproduced by fibroblasts and/or pericytes (58, 59). There are two important classes ofmacromolecules that make up the ECM. The first group are the proteoglycans thatcontain glycosaminoglycans (GAGs) and the free GAG, hyaluronan. GAGs tend toadopt extended conformations that occupy a large volume relative to their mass. Thehigh density of negative charges attracts a cloud of cations, such as Na+, causing influxof water into the matrix. Except for hyaluronan, all GAGs are found covalently attachedto proteins in the form of proteoglycans. Proteoglycans and hyaluronan in connectivetissues form a highly hydrated, gel-like ‘ground substance’ in which the structural andadhesive proteins are embedded; the polysaccharide gel resists compressive forces onthe matrix. The aqueous phase of the polysaccharide gel permits rapid diffusion ofnutrients, metabolites, and hormones between blood and the tissue cells (reviewed in(60)).

The second group is formed by the ECM-glycoproteins that can be divided into twofunctional types: a) structural proteins like collagens (61) and elastin or b) other proteinslike fibronectin and laminin. The collagen fibers both strengthen and help to organize

12

the ECM and rubber-like elastin fibers provide resilience. ECM-glycoproteins such asfibronectins promote the attachment and migration of fibroblasts and various other cellsto the matrix, while laminins promote the attachment of epithelial cells to the basallamina (62, 63).

The major connective tissue blocks of the interstitial tumor stroma includeinterstitial collagens (mainly types I and III), fibrin, fibronectin, elastin and laminin, andGAGs. (57, 64). Although the same building components are used, tumors can vary alot from each other with regard to amounts of ECM. In some desmoplastic tumors, theECM may constitute 90% of the total tumor mass whereas others have only minimalECM comprising 10% or less of the tumor mass. Differences in tumor stroma may alsobe qualitative. Desmoplastic carcinomas contain a lot of dense scirrhous collagen,whereas medulla carcinomas of the breast include a prominent inflammatory cellcomponent with relatively little connective tissue. The composition of the tumor ECMis not static and can change extensively with time (26).

Basement membranes in tumor stromaBasement membranes are important for tissue compartmentalization by acting as filtersand barriers to cell penetration. Basement membranes are formed through specific self-assembly mechanisms and several of their constituent ligands interact with cellularreceptors. The major constituents of basement membranes are collagen IV, laminins,proteoglycans like perlecan and nidogen (entactin) (62). Two networks can be found ina basement membrane. One is formed from collagen type IV and one from differentlaminin isoforms. The collagen IV network is highly crosslinked and maintainsmechanical stability. Laminin networks are probably more dynamic. The two networksare interconnected by the binding protein nidogen, which together with othercomponents, stabilizes the network structures and adds special properties (62, 65).

In tumors, the disruption of basement membranes during tumor invasion reflects ashift in basement membrane turnover, in the direction towards increased degradation.The same tumor cells in a different environment, such as a developing metastasis, mayreturn to basement membrane deposition. Basement membranes in tumors are depositedby stromal cells or by the concerted action of stromal and epithelial cells has beendemonstrated in studies of human carcinomas xenografted to athymic mice (66). A shiftin the direction of increased basement membrane deposition is usually correlated with afavorable prognosis and reflects a high degree of tumor cell differentiation or acompetent host reaction or both (56).

13

Cellular components in tumor stromaBesides ECM molecules and basement membrane components is tumor stromaconstituted of several cellular components. The role of inflammatory cells in tumors ispoorly understood and also contradictory. One way of thinking is that inflammatorycells are recruited to the tumor site to defend the host, another way of thinking is thatthe tumor cells recruit and activate inflammatory cells, and use the assorted array ofsecreted cytokines, cytotoxic mediators, proteases and soluble mediators of cell killing,

such as tumor necrosis factor-α (TNF-α), interleukins and interferons for their ownpurposes like invasion and angiogenesis (74-76).

epithelialcells

quiescentstromal

cells

initiatedepithelium

Activated stroma,inflammatory cells(PMNs, Mø, MCs)

benign carcinoma

Stromal Remodeling(inflammation, angiogenesis,ECM synthesis/degradation)

Malignantconversion

initiation

Promotion

(mutations)

(tumor promoter/wounding)

(growth)(Induction ofinflammation/angiogenesis)

Figure 1. Epithelial - stromal cross talk during neoplastic progression. Epithelialneoplasia is initiated with mutational events; however these cells remain dormant untilwounding or tumor promoters activate the resting stromal cells. These events recruit andactivate inflammatory cells such as PMNs, macrophages (Mφ) or mast cells (MCs). In turn, theactivated stromal compartment stimulates growth and progression of the epithelial cells to forma benign tumor. These cells further activate inflammation and more angiogenesis and induceextracellular matrix remodeling. This altered microenvironment further destabilizes theepithelial cells to undergo malignant conversion to full carcinoma and facilitates metastasis.This model is adopted from Coussens and Werb, 2001.

When growing tumor cells form a lump of cells, they need a supportive stroma andblood vessels to provide them with the right microenvironment, nutrients and oxygen.Fibroblasts produce most of the ECM macromolecules and they probably form a

14

population of cells derived from variable mesenchymal origin. Evidence has beenprovided demonstrating that vascular pericytes have been suggested to be a progenitorcell for matrix producing mesenchymal cells (58, 59, 67-69). Pericytes are cells situatedin the basement membranes outside the vascular endothelium of capillaries and venules(70-73).

A functional lymphatic system has not been demonstrated in tumors (77). It hasbeen proposed that this reflects the collapse of lymphatics within the tumor due tomechanical stress generated by proliferating cancer cells, and that this in turncontributes to increased interstitial fluid pressure (IFP) within the tumor interstitium(78). Although the lack of intra-tumoral lymphatics appears to be a consistent feature,dilated lymphatic vessels are present in peritumoral stroma, occasionally penetratinginto the tumor periphery (78, 79). Metastatic spread of tumor cells via the lymphatics isa common pathway of initial dissemination. It is hypothesized that tumor cells lyingfree in the periductal spaces will be washed with the tide of tissue fluid into thelymphatic system (77).

Collagen type I

Structure

All collagen molecules consist of three polypeptide chains, called α chains that areintertwined into the characteristically triple helix. In each of the polypeptide chains

every third amino acid is glycine, and thus the sequence of a α chain or a collagendomain in a protein can be expressed as (Gly – X – Y)n, where X and Y represent aminoacids other than glycine and n varies according to the collagen type and domain. Thepresence of glycine, the smallest amino acid, in every third position is essential, becauselarger amino acids would not fit into the restricted space in the center of the triple helixwhere the three chains come together. Proline is frequently found in the X position and4-hydroxyproline in the Y position, these two-ring amino acids provide stability for thetriple helix. Hydrogen bonds and water bridges further stabilize the triple helix. Theserequire the presence of 4-hydroxyproline residues, which therefore play an essential rolein the stability of the triple helix. The resulting aggregation of fibrils is characterized bya marked resistance to tensile strength, which is a desired feature in tendons andligaments in order to resist the forces applied to these tissues (reviewed in: (80, 81)).

Decorin (82), fibromodulin (83) and lumican (84), all members of the leucine-richrepeat (LRR) glycoproteins/proteoglycans family bind to fibrillar collagen, therebyretarding the rate of fibril formation and leading to the synthesis of thinner fibrils (85).

15

They are thought to play a role in the regulation of the thickness of the collagen fibrils,providing the right composition and/or structure of a particular tissue. Another group ofproteins that most likely are involved in collagen fibrillar organization are thethrombospondins (86).

BiosynthesisCollagen synthesis involves a number of post-translational modifications. Collagen I isfirst synthesized as a procollagen molecule that has propeptide extensions at both theirN- and C- terminal ends. The main intracellular steps in the assembly of a procollagen

molecule from pro-α chains are cleavage of the signal peptides, hydroxylation of certainproline and lysine residues to 4-hydroxyproline, 3-hydroxyproline and hydroxylysine,glycosylation of some of the hydroxylysine residues to galactosylhydroxylysine andglucosylgalactosyl-hydroxylysine, glycosylation of certain asparagine residues in one orboth of the propeptides and association of the C propeptides through a process directedby their structures (87-92). After the C propeptides have become associated and about

100 4-hydroxyproline residues have been formed in each of the pro-α chains, a nucleusof the triple helix forms in the C-terminal region, and the triple helix is then propagatedtowards the N-terminus in a zipper-like fashion (91, 93).

The procollagen molecules are transported from the endoplasmic reticulum acrossthe Golgi stacks without ever leaving the lumen of the Golgi cisternae (94). During thistransport, the molecules begin to aggregate laterally, and increased condensation of theaggregates results in the formation of granules ready for secretion (94). Theextracellular steps in biosynthesis include cleavage of the N and C propeptides (95),self-assembly of the collagen molecules into fibrils by nucleation and propagation (91,96, 97) and formation of covalent cross-links (98). Collagen synthesis also involves aspecific molecular chaperone, Hsp47 (99, 100), which binds to regions of the triplehelix that are of low stability and thus helps to stabilize the molecules within theendoplasmic reticulum (101). Hsp47 is clearly required for normal development, as ahomozygous knock-out of the Hsp47 gene in mice leads to embryonic lethality (102).

Regulation of biosynthesisCollagen biosynthesis has been shown to be regulated at several levels, includingpromoter activity/transcription, mRNA stability and processing and translational events.

TGF-β is a well-known stimulator of collagen synthesis at the transcriptional level.

Cloning of the promoters for the pro α1 and α2 collagen type I genes has led to the

identification of several elements that confer response to TGF-β, including an NF-1 site

upstream of the mouse pro-α2(I) collagen promoter (103) and a TGF-β response

16

element containing an Sp1-binding site within the human pro-α2(I) collagen I promoter

(104). The pro-α1(I) collagen I gene also contains a TGF-β responsive element with an

Sp1 binding site (105). There are also several reports claiming that TGF-β increasescollagen type I mRNA stability (106, 107). Less is known about regulation of collagen

I mRNA translation. TGF-β increases pro-collagen α1(I) mRNA but not collagen pro-

α2(I) mRNA in lung fibroblasts (108). In vitro translation of RNA yielded however an

increase in the amount of pro-collagen α2(I) in stimulated cells, showing that TGF-β

increased the translatability of thepro-α2(I) transcript maybe via induction of a TGF-βregulatable factor, which controls translation of type I collagen mRNAs (108). Papers Iand II of this thesis also provide evidence for the latter, namely that growth factorsproduced by tumor cells are able to stimulate procollagen I mRNA translation.

Thyroid cancer

Thyroid neoplasms are uncommon, constituting approximately 1% of all cancers (109).Thyroid carcinoma affect more women than man. The median age onset between thefourth and the fifth decades of life. The cause of thyroid carcinoma is not entirely clear.Exposure to radiation is a major risk factor (110). The thyroid gland contains two majorparenchymal cell types, follicular cells and parafollicular cells (also called C-cells).Tumors originating from the C-cells are called medullary carcinomas (111) and will notbe further discussed in this section. The function of the thyroid follicular cells is toproduce thyroid hormones (tri-iodo-thyronine (T3) and thyroxin (T4)). Tumors derivedfrom the thyroid follicular cells are generally divided into two main groups based onhistological appearance and thyroid-specific protein expression, 1) differentiated(papillary and follicular carcinomas) and 2) undifferentiated (anaplastic) carcinomas(see below) (112).

Differentiated thyroid carcinomasPapillary carcinoma. A differentiated tumor type that often grows in papillary-likestructures. It is the most common of all thyroid malignancies (approximately 60 – 80%of all cases). Papillary carcinomas usually affect patients in their third or fourth decadeof life. The tumor often grows locally invasive, but frequently metastasize via thelymphatic drain to regional lymph nodes. In general, this tumor is the slowest growingof all thyroid carcinomas and is also associated with the best prognosis. In a subset ofpapillary thyroid cancer RET proto-oncogene rearrangements are causative events in the

17

pathogenesis (113). However, expression of the RET oncogene has no influence on thelong-term outcome of the cancer (114).

Follicular carcinoma. It is a highly differentiated tumor type that often resemblesthe normal thyroid epithelium. This tumor usually affects patients that are in the fifthdecade of their life, and constitute approximately 20% of all thyroid carcinomas. Itgrows slowly, invading locally, and is frequently detected before metastases occur. Thefollicular carcinomas often give distant metastases, primarily to the lung and bone. Ithas a good prognosis, with an 80% survival rate 10 years after diagnosis (112, 115,116).

Undifferentiated (anaplastic) thyroid carcinoma (ATC)ATC is the ultimate dedifferentiation of the thyroid carcinomas, and as the nameindicate, usually lack most of the thyroid-specific gene expression. It representsapproximately 5% of all thyroid cancers. Most patients affected by ATC are elderlywith a background of preexisting goiter or differentiated carcinoma. The presentingsymptoms is usually a rapidly growing neck mass, resulting in signs of compressionsuch as hoarseness, dyspnea, and dysphagia (116). Tumor morphology shows a solid,trabecular or scirrhous pattern (117) and the tumor invades rapidly adjacent structuresand metastasizes extensively throughout the body. Anaplastic carcinoma is consideredto be one of the most aggressive tumors in man with a mean survival time of about 6-7months following diagnosis (7, 118, 119). ATC do not take up radioactive iodide,therefore as treatment usually radical surgery, irradiation and chemotherapy are used(116)

TGF-β and TGF-β signaling

TGF-β superfamily of ligands

In human nearly thirty members of the TGF-β superfamily have been describe. They

belong to a family of dimeric polypeptide secreted factors that include the TGF-βs, thebone morphogenetic proteins (BMPs), inhibins and activins (120). There are three

isoforms of TGF-β: TGF-β1, TGF-β2 and TGF-β3 (reviewed in (121), having similar

biological effects. Each TGF-β isoform is encoded by a distinct gene and is expressed in

a tissue specific fashion. The TGF-β1-null mice die in utero because of defects invasculogenesis and hematopoiesis or die shortly after birth because of overwhelming

inflammatory infiltration of multiple organs; TGF-β2- and -β3-null mice have multiple,

18

although different and non-overlapping, developmental abnormalities (reviewed in:(122)).

TGF-β is synthesized as a part of a large precursor molecule containing a

propeptide region in addition to TGF-β (123). After it has been secreted, most TGF-β is

stored in the ECM as a complex of TGF-β, the propeptide, and a protein called latent

TGF-β binding protein (LTBP) (123). The attachment of TGF-β to LTBP prevents it

from binding to its receptor. Latent TGF-β can be activated through multiple

mechanisms potentially involving integrin-αvβ6, mannose-6-phosphate receptors,plasmin, MMPs 2 and 9 and thrombospondin-1 (TSP-1), a glycoprotein secreted bymost cells and incorporated into the extracellular matrix (123-125)

TGF-β superfamily receptors

TGF-β and related factors signal through a family with transmembrane proteinserine/threonine kinase receptors (126). These receptor proteins can be divided into two

main groups, the type I and II receptors (TβR-I and TβR-II). Accessory receptors like

TβR-III and endoglin are able to bind TGF-β and present it to TβR-II, but are not

directly involved in TGF-β signaling (127). Ligand binding to the constitutively active

TβR-II results in the formation of a TβR-I/TβR-II heterotetrameric complex (Figure 2)

(128, 129). The TβR-II kinase phosphorylates TβR-I in the GS-box resulting in

activation of TβR-I kinase. Activated TβR-I recruits and phosphorylates intracellularsignaling molecules called Smads (128, 129).

The Smad proteins involved in TGF-β signaling are Smad2, Smad3, Smad4 and

Smad7. Activated TβR-I binds Smad2 and Smad3, the receptor activated Smads (R-Smads). These phosphorylated Smads then form heteromeric complexes with the co-Smad, Smad4, which translocates to the nucleus, (130, 131). The inhibitory Smad (I-Smad), Smad7, lacks the carboxy-terminal phosphorylation motif of Smad2 and Smad3

but interacts with TβR-I and the R-Smads, inhibiting the pathway (132, 133). Smad complexes, once translocated into the nucleus, can regulate transcription in

several ways. The MH1 domains of Smad3 and Smad4 can bind directly to DNA via aSmad binding element. The MH-2 domains of Smads have a low affinity for theirbinding sequences, therefore they cooperate (in particular Smad2, which cannot itselfbind to DNA) with other transcription factors, like AP-1, Sp1 and TFE3, or they interactwith transcriptional coactivators, including CBP and p300, or corepressors like c-Skiand SnoN for both activation and inhibition of transcription (reviewed in (122)).

Recently a number of inhibiting proteins are recognized that interact with thereceptors or prevent receptor phosphorylation. The immunophilin FKBP12 binds to the

GS domain of the TβR-I, inhibiting its signaling function (134). BAMBI is a

19

transmembrane protein, which is related to TGF-β family receptors but lacks anintracellular kinase domain (135). STRAP binds Smad7 and recruits it to the activated

TβR-I, preventing Smad2 and Smad3 access to the receptor (136). TRAP-1 is an

intracellular protein that binds activated TβR-I and prevents downstream signaling(137). SARA (smad anchor for receptor activation) is involved in the stimulation of

TGF-β signaling, recruiting Smad2 to the active TβR (138).

PP

TβR-II TβR-I

FKBP 12

BAMBI

PP

TβR-II TβR-I

PP

TGF-β TGF-β

SARA

Cell membrane

Smad2/3

Smad2/3

TRAP1

STRAPSmad7

Smad 4

Nucleus

P

P

PSmad 4Smad2/3

DNA-bindingCofactor

Co-activator/Co-repressor

P

PSmad 4Smad2/3

Transcriptional complex

Smad2/3

Smad2/3

Figure 2. Schematic illustration of the TGF-β signaling pathway

20

TGF-β inhibition of the cell cycle

TGF-β induces growth arrest in most epithelial cells, including many carcinomas in

culture (139, 140). Stimulation of six thyroid carcinomas with TGF-β resulted in growthinhibition in a majority of the thyroid carcinoma cells (141) and similar results were

obtained after TGF-β stimulation of normal porcine thyroid follicle cells (142). TGF-βis generally mitogenic to mesenchymally derived cells through the induction ofmitogenic growth factors such as PDGF (143).

Proliferating cells go through different stages of the cell cycle. In the G0 phase,cells are resting and can be withdrawn from the cell cycle. When receiving mitogenicsignals, cells will enter the cell cycle. As shown in Figure 3, each phase of the cell cycleis controlled by specific cyclins and their respective kinases, cyclin dependent kinases(cdks). Progression through the cell cycle involves the phosphorylation of theretinoblastoma tumor suppressor gene product pRB by D and E cyclin: cdk kinaseactivities. The cyclins bind to pRB and related proteins (p130 and p107). Loss of pRB, atumor suppressor gene, promotes oncogenic transformation (reviewed in: (144)).

M

G1

STGF-β

p15

p27(redistribution)

p21

G2

RB- P

P - RB- P

PP - RB- P

PP - RB- P

P

PP - RB- P

P

Cycl A

Cycl E

Cycl D

cdk 2

cdk 2

cdk 4/ 6

G0

Figure 3. TGF-β effects on the cell cycle.This drawing is based on a similar drawing made by Leslie Gold (1999).

The formation of cyclin/cdk complexes is regulated by the cyclin-activating kinase(CAK) (139). Cell cycle arrest is associated with hypo-phosphorylation of pRB and a

21

decrease in G1 cyclins and cdks. Hypo-phosphorylated pRB binds to E2F complexesthat are released when pRB is phosphorylated. Released E2F complexes transactivatetarget genes, such as c-myc that control the expression of genes responsible for cellcycle progression (144).

TGF-β has been shown to inhibit the mRNA and protein expression/accumulationof most cdks and cyclins and also prevent cdk/cyclin complex formation (139, 145).

The cell cycle arrest by TGF-β is also mediated by induction of cylin-dependent kinaseinhibitors (CKIs) that bind to the cyclin-cdk complexes and inhibit their ability tophosphorylate pRB, thus maintaining cells in G1. There are two groups of mammalianCKIs: 1. The ink4 group, including p15, p16, p18 and p19, interferes specifically withcyclin D binding to cdk4/6 and 2. The Kip/Cip proteins (p21, p27 and p57) that inhibit

the cyclins A, D and E and their respective cdks (146, 147). TGF-β specifically inducesan increase in p15, p21 and p 27 protein levels. The p15 protein binds to the cyclin Dcomplexes, inactivates the complex, and releases p27 from the cyclin D complexes sothat it can subsequently bind cyclin E-cdk2, preventing progression to S phase (145,146).

PDGF and PDGF signaling

PDGF isoformsPlatelet-derived growth factor (PDGF) is a major mitogen for cultured fibroblasts,smooth muscle cells, and other mesenchymally derived cells (reviewed in: (50)).PDGFs form a family of heterodimeric or homodimeric chains that are synthesized asprecursor molecules undergoing proteolytic maturation. The synthesis of PDGFisoforms is carefully regulated, and their action on receptors is modulated by interactionwith components in the matrix as well as with soluble binding proteins. PDGF isoformsare dimeric molecules; they bind two receptors simultaneously and dimerize receptorsupon binding (148-150).

There are four isoforms of PDGF; PDGF A, B, C and D forming homo- andheterodimers (Figure 4). PDGF-AA, AB and BB were identified more than 15 years ago(151, 152) whereas the homodimers PDGF-CC and PDGF-DD were only recentlydiscovered recently (153-155). Two PDGF receptor chains have been described, the

PDGF-α- and β- receptor chain. The α-receptor binds the A, B and C chains of PDGF

with high affinity, whereas the β-receptor binds the B and the D chain with high

affinity. However in cells expressing both PDGFR-α and -β, PDGF-D activates both

receptors, probably resulting from PDGFR-α/β heterodimerization (154).

22

PDGF-CC PDGF-AA PDGF-AB PDGF-BB

PDGFR-αα

PDGFR-ββPDGFR-αβ Receptors

PDGF-isoforms

Receptor

PDGF-DD

Figure 4. The PDGF family.

PDGF receptorsUpon ligand binding, PDGF receptor units dimerize, autophosphorylate on tyrosineresidues, and several Src homology 2 (SH2) domain-containing signal relay enzymesincluding Src family tyrosine kinases (pp60cc-src, p62c-yes, p59fyn), phosphatidylinositol-

3-kinase (PI3K), phospholipase Cγ1 (PLCγ), the phosphotyrosine phosphatase SHP-2,and the adapter molecules Shc, Nck and Grb2, stably associate with the tyrosinephosphorylated receptor (156, 157). Some of the signaling molecules interact with theactivated receptor at unique binding sites. Others associate with the receptor in a more

loose way meaning that they can be recruited to several binding sites. The PDGFR α-

and β-subtypes initiate distinct signaling pathways, which results in differences inbiological responses. Signaling is modulated positively and negatively, extracellularlythrough interaction with matrix molecules and intracellularly through cross talk withdifferent signaling cascades (reviewed in: (50).

23

The biological activities of PDGF and TGF-β

The role of PDGF and TGF-β in fibrosis.

PDGF and TGF-β play an important role in stimulation of connective tissue in vivo,overactivity has been implicated in the development of fibrosis in various organ systems(50, 158). Fibrosis is often associated with an inflammatory reaction, wherein macrophages

can be an important source of PDGF and other factors (41, 51, 159). TGF-β is known todirectly upregulate the fibrillar and, to a lesser extend, non-fibrillar collagens, otherECM components including fibronectin and tenascin, the basement membranecomponents laminin and entactin, and proteoglycans including perlecan (160). PDGFscan, besides ECM induction also stimulate proliferation of fibroblasts and this enhances

the fibrotic reaction in two ways (161). More over TGF-β is able to increase its own

expression and generate potent autocrine loops (162). The combination of TGF-βoverproduction and highly proliferative cells is characteristic of some fibroticconditions and may be essential to the development of chronic, progressive fibrosis.

TGF-β regulates the expression of many proteins responsible for matrix degradation.

Although the effect of TGF-β on MMPs is mixed, it upregulates protease inhibitors,such as TIMPs (163) and PAI-1 (21). This results in a decrease in plasminogenconversion to plasmin, a protease that degrades certain ECM proteins and in additionactivates MMPs (163, 164).

The role of PDGF and TGF-β in wound healing

TGF-β and PDGF act on several cell types involved in wound healing. Wound repair isa tightly regulated process, which involves the inflammatory response, proliferation offibroblasts and transformation of fibroblasts into myofibroblasts, and extracellular

matrix synthesis. Platelets store both TGF-β and PDGF, which are released upon injury

(165, 166), and activated macrophages (167) and lymphocytes (168) produce TGF-β.

TGF-β1 and PDGF are potent chemoattractants for monocytes (169) and fibroblasts(35, 170) and may promote influx of these cells into the site of injury. PDGF stimulates

these macrophages to produce and secrete other growth factors, including TGF-β, thatare important for various phases in the healing process and PDGF stimulates productionof several matrix molecules, like fibronectin, collagen, proteoglycans, and hyaluronicacid (reviewed in (50)). PDGF also stimulates secretion of collagenase by fibroblasts(171) and has been shown to stimulate collagen gel contraction (172) suggesting a rolein wound remodeling and wound contraction. Wounds treated with PDGF showed anincrease in granulation tissue rich in fibroblasts and GAGs and an increased rate of re-

24

epithelialization and of neovascularization (173, 174). Exogenous PDGF does not alter

wound repair but increases its rate. Exogenous administration of TGF-β1 to woundedanimals stimulates collagen fibrillogenesis at the wound site (174, 175), thereby

promoting wound healing. However, persistent expression of TGF-β1 induces tissuefibrosis.

The role of PDGF and TGF-β in cancerMany tumors express PDGF and cognate receptors, and an autocrine stimulation oftumor cells can be assumed. In human glioblastoma, the expression pattern of PDGFand its receptor provides evidence for two autocrine loops: PDGF-A chain and the

PDGFR-α are present in tumor cells, whereas PDGF-B chain and PDGFR-β are highlyexpressed in the stromal compartment of the tumor (176-178). Overexpression of

PDGFR-α and PDGF-A was particularly seen in glioma cells of high-grade tumors,

suggesting that autocrine activation of the PDGFR-α may drive the proliferation of

glioma cells in vivo (179). In contrast PDGF-B and PDGFR-β seem to promote tumorgrowth by activation of the stromal compartment and stimulation of angiogenesisleading to vascularization of the tumor (180).

Further evidence for an involvement of PDGF and PDGF receptors intumorigenesis has been provided by findings of structural aberrations of thecorresponding genes that lead to overexpression or expression of an abnormal protein.

In a few cases of glioblastoma, the PDGFR-α gene is amplified (181-183), leading toreceptor overexpression. In cases of dermatofibrosarcoma protuberans and giant-cellsarcoma, the PDGF-B chain gene is rearranged by a chromosomal rearrangement (184).

An additional role for PDGF-BB has been suggested in epithelial tumor formation,implicating that it does not only promote tumor growth by inducing mesenchymal cellproliferation and angiogenesis, but the activated stromal environment may also inducetumorigenic conversion of epithelial cells via a paracrine mechanism (185). Theimportance of PDGF signaling in tumorigenesis seems not to be limited to neoplasms ofmesenchymal origin.

Most cancers have mutations/deletions disabling a component of the TGF-β

signaling pathway. Most of these mutations occur in the basic TGF-β signaling pathwayand are usually associated with progression rather than with initiation of malignancy. In

several types of cancer TβRI and TβRII protein levels are down regulated and this isusually correlated with a decreased patient survival (186). Smad4 is inactivated in alarge percentage of pancreatic and colorectal cancers and is correlated with invasion andmetastasis (187).

25

Although TGF-β is a potent growth inhibitor in epithelial tissues, it is both asuppressor and a promoter of tumorigenesis (188). Experiments in tumor bearing mice

provided evidence for a role of TGF-β in suppression of tumor development, during the

early stages of tumor progression. TGF-β1 heterozygous null mice show increasedhepatocyte proliferation, and decreased apoptosis in the lung and liver. Whenchallenged with carcinogens, these mice develop liver and lung tumors of greater size,

number and malignant potential than the controls, suggesting a role for TGF-β1 intumor suppression (189).

In later stages of tumor progression high levels of TGF-β expression are correlated

with advanced clinical stage of the tumor (190). High TGF-β levels expressed by tumorcells could contribute to tumor growth indirectly by suppression of the infiltrating

immune cells or stimulating angiogenic factors (191). However TGF-β can also actdirectly on cancer cells to foster tumorigenesis. Tumor cells that have selectively lost

their growth-inhibitory responsiveness to TGF-β but retain an otherwise functional

TGF-β signaling pathway may exhibit enhanced migration and invasive behavior in

response to TGF-β stimulation (192-194).

TGF-β can promote tumor cell metastasis in many different ways. Of interest is the

ability of TGF-β to induce an epithelial to mesenchymal transition (EMT) (193). EMTis characterized by the down regulation of proteins involved in cell-cell adhesion andupregulation of molecules important for cell-ECM associations, ultimately leading toenhanced migratory and invasive properties of the cell. A switch from an epithelial tofibroblastoid phenotype occurs frequently during late stages of carcinoma progressionand correlates with the metastatic potential of tumor cells. These observations are stillcontroversial, since the majority of this evidence is derived from experimentalmetastasis assays that utilize engineered carcinoma cell lines.

Interstitial fluid pressure

At all times fluid filtrates through the capillary wall out into the interstitium. Accordingto the Starling hypothesis the rate of fluid filtration can be described by the equation JV

= �PK, where JV is filtration rate, P is differences in fluid pressures and K is a constant

expressing capillary area and permeability (for a review see: (195)). The differences inpressure, P, is generated by differences in the colloid osmotic pressures in plasma(COPPL) and in the interstitial fluid (COPIF), and between capillary hydrostatic pressure

(CHP) and interstitial fluid pressure (IFP) as can be described in the equation P =

∆

∆

∆

∆

26

(CHP - IFP) - σ (COPPL - COPIF), where σ is a plasma protein reflection coefficient.

The commonly accepted role for the IFP has been to maintain constant interstitialvolume and transcapillary fluid flux. An increase in transcapillary filtration will raiseinterstitial volume and thereby IFP which will act across the capillary to limit furtherfiltration, as well as enhance lymph drainage from the tissues (195). The IFP is not a"passive controller" of the interstitial fluid volume, it has been observed that interstitialpressure fell as low as -150 mmHg concomitant with edema formation in burn injuriesto the skin. Thus, interstitial pressure actually becomes the driving force for the edemaformation rather than being a "passive controller" of the interstitial fluid volume.

An active role for connective tissue cells in the maintenance of IFP is suggested by

the finding that a subdermal injection of anti-β1-integrin IgG results in rapid edema

formation, including an increased negativity in IFP (196). Monoclonal antibodies to the

collagen-binding integrin α2β1 also induce a lowering of IFP in rat dermis, and thiseffect is counter-acted by PDGF (197). PDGF can also normalize a lowering of IFPinduced by anaphylaxis in normal rat and mouse dermis (197, 198). Furthermore,prostaglandins are potent modulators of IFP (199). Fibroblasts that are cultured in athree-dimensional collagen lattice will contract the lattice in a process that require

several hours (200, 201). β1 integrins are instrumental for fibroblast-mediated collagengel contraction. and are stimulated by PDGF (172). The data also demonstrated that

PDGF increased the apparent avidity of the collagen-binding β1 integrins in this assaysystem (172). A similarity between the process of collagen-gel contraction in vitro andcontrol of IFP and edema formation in vivo has been reported (196, 197, 199, 202, 203).From these and other data a model for how connective tissue cells control IFP has beendeveloped. According to this model connective tissue cells exert tension on a

collagen/microfibril network, via the collagen binding integrin α2β1 (196, 197, 204,

205). These tensile forces control the intrinsic swelling properties of thehyaluronan/proteoglycan ground substance in ECM (206). If the connective tissue cellslowers their tension the tissue is allowed to swell, which is reflected in a lowering ofIFP.

Tumor interstitial fluid pressure

In most solid tumors the normal fluid balance is disturbed and a pathologically elevatedtumor interstitial fluid pressure (TIFP) is recorded (207-209). It has been suggested thatan elevated TIFP forms a barrier for the delivery of anti-cancer drugs into the tumor(210, 211), therefore lowering of the TIFP is thought to be a useful therapeutical target.

27

It is not clear what causes the high in TIFP but several mechanisms have been proposed.Jain and co-workers claim that TIFP is high because of high capillary permeability(leaky vessels) and impaired lymphatic drainage (78, 212, 213). Other suggestions arethat neoplastic cell proliferation increases vascular resistance and TIFP (214). and themechanical stiffness of the tumor tissue may inhibit macromolecule transport (215). Afinal mechanism could be that the reactive stroma keeps the tumor stroma in aconstantly contracted state (197, 198).

Several agents that induce a lowering of TIFP have been identified, including

nicotinamide (216), tumor necrosis factor α (217) and dexamethasone (218). It has,however, not been established whether these agents increase the uptake of drugs into thetumor. Tumors treated locally with Prostaglandin E 1 methyl ester (219) demonstrated

an acute lowering in TIFP whereas systemic treatment with specific inhibitors forPDGF signaling resulted in a chronic lowering of TIFP (220). In both these studies, thelowering in TIFP is followed by an increased capillary to interstitium transport of thelow molecular weight compound 51Cr-EDTA. PDGF is hypothesized to increase thetension, exerted on the ECM-fiber network by stimulating connective tissue cells (197,198) and keep the tumor in a contracted state.

In the present work, the effects of a specific TGF-β1 and -β3 inhibitor on TIFP(paper III and IV) were investigated. Longterm systemic treatment resulted in asignificant decrease of TIPF and reduced tumor growth.

28

Aims

The overall aim of the present study was to increase the understanding of mechanismsfor tumor stroma formation and physiological functions.

The specific aims were:

• To characterize human anaplastic thyroid carcinoma (ATC) tissue specimens andcell lines with regard to interstitial collagen production.

• To establish an experimental animal model for ATC that allow for the study ofstromal reactions.

• To investigate the role of TGF-β for stromal collagen deposition and hydraulicproperties, using the experimental ATC model.

• To investigate the relation between stromal collagen deposition and tumoralinterstitial fluid pressure in experimental ATC model.

29

Methodology

Details of the methods used in this thesis are described in the respective papers andmanuscripts. In this section only specific key methods are discussed.

ATC cell lines (paper I, II, III and IV)Six cell lines, derived from different human ATC tumors were used in the experimentalstudies of tumor - stroma interactions. As a representative cultured cell to study thestroma responses, we choose the well-characterized human diploid foreskin fibroblastAG1518 with low passage numbers. We are aware of the fact that this type offibroblasts are not present in tumors, but in vitro models are artificial models and thefindings from in vitro experiments can give important indications for the in vivo model,but do not represent the in vivo situation.

Quantification of collagen synthesis (paper I, II and IV)Collagen synthesis was quantitated using an assay that determines the production ofmetabolically labeled triple helical collagens.

Cells or co-cultures of cells were plated in 96-well plates. At least, quadruplicate

samples were incubated for 5 h at 37°C prior to a 18-20 h. labeling period. After thepulse period the whole cell cultures were digested by pepsin, followed by precipitationin high salt at both low and neutral pH. The samples were separated on polyacrylamidegel electrophoresis in sodium dodecyl sulphate (SDS-PAGE) and the metabolicallylabeled collagen bands were quantified using a Phosphor Imager (Fuji, Tokyo, Japan).

Collagen can also be quantified by other methods, for example by collagen Westernblot, using anti collagen antibodies. Since we detect metabolically labeled collagenbands, synthesis can be quantified directly, whereas this would not have been possibleusing collagen Western blot. The disadvantages of the present method are: 1, the risk ofloosing collagen protein during the various precipitation and washing steps; 2, amethod, which is quite laborious and 3, the costs for working with radioactivity.

Cell cultures on poly[HEMA] (paper II)Poly-(2-hydroxyethyl methacrylate) (poly[HEMA]) is an acrylate, used to coat cellculture plastic. Cells plated on these poly[HEMA] films attach, but are refrained fromspreading (221) and collagen I protein synthesis is reduced dramatically in fibroblasts

cultured on poly[HEMA]. The pro-α1(I) collagen mRNA expression level was,

however, not affected. Stimulation of these cells with PDGF-BB and TGF-β inducescollagen I synthesis to almost 100%, as when the cells were allowed to spread but does

30

not affect collagen mRNA expression levels. (222). This method does thus allow for the

study of pro-α1(I) collagen mRNA translation.Cells and co-cultures were plated in 96-well plates coated with an ultra-thin film of

poly[HEMA]. Subsequently cells were seeded on top of the poly[HEMA] film inserum-free medium and stimulated with growth factors followed by a [14C]Glycinepulse for 18-20 h.

One drawback with this method, and possible source of variation is that the quality ofthe coated poly[HEMA] films differs between experiments. Therefore, it is important toadd to each experiment the proper controls and to monitore the quality of eachexperiment. Also cell spreading has to be assessed microscopically in each experiment.The use of poly[HEMA] coated plates allows studying mRNA translation with an intacttranscription machinery and the use of the same experimental make-up as withoutpoly[HEMA].

Immunohistochemistry and in situ hybridization (paper I, II, III and IV)Death by apoptosis is characterized by membrane inversion, exposure ofphosphatidylserine residues, blebbing, fragmentation of the nucleus, chromatincondensation and DNA degradation. The TumorTACS in situ apoptosis kit detectsapoptotic cells by the incorporation of biotinylated nucleotides onto the free 3’-hydroxylresidues of fragmented DNA.

Apoptosis staining was performed with a few adaptations. In brief, frozen sectionswere subjected to proteinase K followed by a second fixation. Fragmented DNA waslabeled with biotinylated nucleotides and the staining was visualized using the ABC-DAB kit.

Critical steps in the assay are: 1) the fixation steps to prevent the loss of lowmolecular weight DNA fragments and to keep the tissue morphology; 2) the proteinaseK treatment to make the DNA accessible to the labeling enzymes; 3) quenching ofendogenous peroxidase by hydrogen peroxide; 4) labeling of the DNA fragments withbiotinylated nucleotides by the (TdT). The results of positive control specimen canreveal whether the staining was successful or not.

In situ hybridization is a well established method to detect collagen type I mRNA intumor tissue sections. DIG-labeled probes increasingly replace the radioactive labeledprobes and allow detecting the mRNA staining with normal light microscopy, as well asvarious double staining protocols.

In situ hybridization was performed with Digoxigenin (DIG)-UTP labeled pro-

α1(I) collagen RNA probes, Sections were treated with proteinase K, prehybridized at

72°C followed probe hybridization at 72°C. An alkaline phosphatase-conjugated anti-

31

DIG antibody detects the hybridized RNA probes, which are then developed withNBT/BCIP. Carcinoma cells were double stained with tissue polypeptide antigen (TPA)which can be used to specifically detect carcinoma cells.

Critical steps in this assay are: 1) probe labeling, 2) proteinase K treatment and 3)

hybridization temperature. Labeling of pro-α1(I) collagen RNA probes was made by invitro transcription using T3 and T7 RNA polymerases. Inefficient labeling might giverise to labeled probe fragments of different sizes that impair the in situ hybridization.Sections treated with proteinase K for extended time periods, lose their morphology.Too short proteinase K treatment prevents access of the probe into the cells. Thehybridization temperature is also of importance. Temperatures below the optimumtemperature decrease hybridization stringency thereby increasing unspecific staining.Temperatures above the optimum might give no staining at all. The samples should notget dry during the hybridization; this increases the risk for unspecific staining as well.

Measurement of TIFP (paper III)TIFP was determined using the wick-in-needle technique (223). This method is widelyused for determinations of tissue interstitial fluid pressure and gives trustworthymeasurements. In skin it is possible to measure the Pif by using thin sharpened glasscapillaries (micropuncture). Unfortunately, this is not possible in tumors because theglass capillaries will break when inserted in the rigid tumor tissue. The 23-gauge needledamages the tumor and will give rise to scar formation; this is a disadvantage especiallywhen TIFP is to be measured more than once in the same tumor.

32

Results

In this thesis we set out to elucidate the mechanism(s), which give rise to the fibroticreaction in human ATC. Tumor – stroma interactions are known to determine the tumorphenotype (11, 224, 225), though the respective roles of the carcinoma and stromal cellswere not known. We also investigated the relevance of tumor stroma composition forone physiological parameter, namely TIFP. This is of clinical interest since recentexperimental data from our laboratories suggest that a lowering of TIFP increasesuptake and efficiency of classical anti-cancer drugs (219, 220).

Paper ITissues and cell lines derived from human ATC were used to study the fibrotic reaction

in ATC. In situ hybridization showed that pro-α1(I) collagen mRNA was expressed

throughout the tumor stroma compartment. In one out of five cases, pro-α1(I) collagenmRNA was also expressed in a subset of carcinoma cells. The latter suggested thatsome carcinoma cells indeed have the ability to synthesize collagen I, but that tumorfibroblasts were the main source of collagen production. Immunofluorescense stainingusing an anti-procollagen type I antibody detecting newly synthesized collagen type Iprotein, revealed synthesis mainly in the stromal cells situated in close vicinity to nestsof tumor cells. This suggested that communication between carcinoma cells and stromal

fibroblasts stimulated translation of pro-α1(I) collagen mRNA.

To study pro-α1(I) collagen synthesis by carcinoma cells, six well-characterizedATC cell lines were used in vitro. All cell lines expressed prolyl 4-hydroxylase mRNAindicating that they are, in theory, able to synthesize triple helical collagen type I. TheATC cell lines lacked a number of their epithelial characteristics, this was emphasizedby the fact that three of the six cell lines synthesized native triple-helical collagen typeI, though to a much lower extent than AG 1518 fibroblasts. A mesenchymal phenotypewas also underscored by the finding that 5 out of 6 ATC cell lines expressed vimentinbesides cytokeratin or that they had lost cytokeratin expression completely.

Taken together, data in this paper showed that stromal fibroblasts are the mainproducers of collagen type I ATC. The carcinoma cells seem to induce the collagen

synthesis in stromal fibroblasts by increasing pro-α1(I) collagen mRNA translation.ATC form a heterogeneous group, in which a subset of the carcinoma cells cancontribute to the fibrotic reaction by synthesizing collagen type I, though only to aminor extent.

33

Paper IIInjection of the six ATC cell lines subcutaneous in the upper left flank of athymic micelead to two different tumorigenic responses. The 3 collagen producing cell lines had as agroup a much lower tumorigenicity (17%) than the non-collagen producing cell lines(60%). The tumors developed from the collagen producing ATC cell lines displayed a

morphology with pro-α1(I) collagen mRNA expressing carcinoma and stromal cellsinterdispersed. The non-collagen producing ATC cell lines gave rise to tumors with adifferent morphology, displaying 2 distinct compartments. Carcinoma cell islets, that

are surrounded by pro-α1(I) collagen mRNA expressing cells confined to a welldelineated stromal compartment.

In this study an in vitro model was also set up to study the effect of stimulatoryfactors on collagen type I mRNA translation. Human diploid foreskin AG 1518fibroblasts and ATC cell line KAT-4 were cultured on poly[HEMA]-coated microtiterplates, that provides a culture condition, which allows cell for attachment but not cellspreading. We investigated our hypothesis that ATC cells induce collagen I synthesis bystimulating collagen mRNA translation. Addition of conditioned media from the sixATC cell lines (paper I) to AG 1518 fibroblasts cultured on poly[HEMA] had no effecton collagen type I synthesis by AG 1518 fibroblasts. Co-cultures of KAT-4 tumor cellsand AG 1518 fibroblasts on poly[HEMA]-coated plates resulted in a stimulation ofcollagen type I synthesis correlated with the number of KAT-4 cells. These dataindicate that cell-cell contact is required for collagen type I stimulation. The use of

specific inhibitors for PDGF-BB and TGF-β1 in KAT-4/ AG 1518 fibroblast systemdemonstrated that these two factors are involved in KAT-4 induced collagen type Isynthesis by fibroblasts. Each inhibitor alone resulted in a partly inhibition whereas thecombination displayed an adding up effect.

Paper IIIThe results from papers I and II, revealed that carcinoma – stroma cell communicationswere involved in the stimulation of collagen I mRNA translation and indicated a role for

both PDGF and TGF-β herein. The availability of a specific TGF-β inhibitor made it

possible to investigate the role for TGF-β in the generation of an elevated TIFP inexperimental xenograft ATC.

Athymic KAT-4 tumor bearing mice were injected intravenously with a soluble

TGF-β receptor type II - murine Fc:IgG2A fusion protein, which specifically inhibits

TGF-β1 and -β3. Inhibitor treatment resulted in a significant lowering of TIFP after 10days of treatment. This decrease was shown to be dose dependent, and was not an acuteeffect. Measurements after1 day of treatment did not display a lowered TIFP. The

34

decrease in TIFP did not lead to edema formation in the tumor, since tumor tissue watercontent was equal in treated and control tumors.

Another remarkable finding was that, during the first 10 days after treatment, theinhibitor treated tumors had increased in size much more than the control tumors.Measurements of apoptosis index in 10 day treated tumors, however, displayedincreased apoptosis in inhibitor treated tumors and also induced expression of the cellcycle inhibitor p27Kip1. These contradictory results indicate that inhibitor treatmentinitiated rapid tumor growth at early time points after the start of treatment, but after 5-10 days of treatment apoptosis was induced and tumor growth strongly reduced.Prolonged inhibitor treatment, with the administration of a second dose after 14 days,

confirmed these findings, showing that TGF-β inhibitor treatment initiates an initialgrowth followed by a reduction in tumor growth after 14 days and control tumorsovertaking the lead in growth after 15-20 days of treatment.

The TGF-β1 and -β3 inhibitor did not affect KAT-4 proliferation in vitro, nor thelevel of phosphorylated Smad2 protein, indicating that KAT-4 carcinoma cells are

unresponsive to TGF-β inhibition. Together the data suggest that TGF-β1 and -β3 maybe interesting targets for novel anti-cancer treatment directed to the stroma, first bylowering TIFP and thereby potentially increase uptake of anti-cancer agents intotumors, and second by its potential role in maintaining a supportive tumor stroma.

Paper IV

In this study we investigated whether the TGF-β1 and -β3 inhibitor affected collagendeposition in experimental ATC formed by the human ATC cell line KAT-4. Sirius redstaining showed no detectable difference in overall collagen deposition in inhibitortreated or control tumors. Further analysis showed that inhibitor treatment did not affect

pro-α1(I) collagen mRNA as investigated by in situ hybridization nor does it change

overall tumor pro-α1(I) collagen mRNA levels as investigated by Northern blot analysisof 10 day control and inhibitor treated tumors.

Determinations of pepsin extractable interstitial collagen proteins demonstrated amarginal reduction after inhibitor treatment. This was confirmed by determinations ofthe triple helical collagen component hydroxyproline (HP) in hydrolyzed tumor tissues,showing a 35% reduction. Hyaluronan (HA) was measured in the same tumorsdisplaying only a marginal increase in inhibitor treated tumors.

The (HP/HA) ratio however, displayed a significant reduction in TGF-β1 and -β3inhibitor treated tumors. A high collagen to hyaluronan ratio may be of importance for

the generation of a high TIFP, since TGF-β1 and -β3 inhibition reduces TIFP in KAT-4tumors.

35

Conclusions and future perspectives

ConclusionsOur studies aimed to elucidate the mechanisms for tumor stroma formation andphysiological functions have resulted in the following conclusions.

• The ATC cells have a dual effect by on one hand, stimulating the synthesis ofcollagen type I protein in the stroma and on the other having the ability to producesmall amounts of collagen type I themselves.

• Collagen producing ATC cell lines display a reduced tumorigenicity upon xenografttransplant and display an interdispersed carcinoma – stroma cell morphology.

• PDGF and TGF-β produced by the carcinoma cells induce pro-α1(I)collagenmRNA translation in fibroblasts via direct cell – cell contact.

• Addition of the TGF-β1 and -β3 inhibitor in an experimental KAT-4 model resultsin a significant decrease of TIFP.

• TIFP lowering seems to be the result of a change in the HP/HA ratio in the tumorstroma.

Future perspectives

We have found that TGF-β1 and -β3 inhibition leads to a lowering of TIFP in theexperimental KAT-4 tumor model. It would be interesting to see if lowering of TIFPresults in an increased uptake of low and high molecular weight anti-cancer drugs. Thiscould be tested by microdialysis with a radioactive compound.

The change in tumor stroma composition that was found upon inhibitor treatmentmay be correlated with changes in other stroma components, such as the small collagenbinding proteoglycans that have a role in fibril formation of collagen and sequestration

of TGF-β.

Treatment with the low dose of TGF-β1 and -β3 inhibitor (0.1 mg/ml) resulted in anincreased TIFP and did not display the fast initial growth immediately after inhibitoradministration; these tumors had grown even less than the control tumors. Thesefindings indicate that the inhibitor might have a biphasic effect depending on dosage.

The initial tumor growth after inhibitor administration is difficult to explain. Mostlikely the inhibitor affects the tumor stroma in such a way that it initiates rapidcarcinoma cell growth. The mechanism behind this is however, presently unclear andneeds to be elucidated.

36

Acknowledgements

Very many thanks to .......

The Swedish Cancer Foundation, The Swedish Medical Research Council, The GustavV:s 80-årsfond, The Åke Wiberg Foundation and The Göran Gustavsson Foundation.

The department of Medical Biochemistry and Microbiology for providing excellentfacilities.

My supervisors for all their help and support.Kristofer Rubin, for teaching me research and sharing with me your vast knowledge.You have been a great help during the hectic period of writing. Thank you also for theregular nice parties at your house.Nils-Erik Heldin, for supporting the progress of the project and all other contributions.You have an excellent eye for structure and lay out, which has been very helpful.

All the members in our group past en present. The oldies for me are: Mikael Ivarsson, Alan McWirther, Bianca Tomasin-JohanssonMaria Branting, Karina Åhlén, Lisa Tung,Therese Dahlman, Nasimeh Khorasani, AnnaKarin Ekwall-Hultgård, Patrick Ring, Marcin Kowanetz and Christian Sundberg,though you’re back again.Special thanks to Mikael and Therese for bringing me into their projects. This is theresult of it!

The present group members: Åsa Lidén, Alexei Salnikov, Gunilla Grundström, EvaAndersson, Cecilia Rydén and Ann-Marie Gustafson.Special thanks to Åsa and Gunilla for all the long good conversations, not only duringlunchtime and for all your help and care.Also special thanks to Ann-Marie and her mousies for helping me with the animal workand for your good sense of humor and laughter.

All the people in corridor D11:3.

Our badminton team, Gunilla, Kia, Sorin, Maria, Gunbjörg. Thanks to Jonas, Mårtenand Staffan O. for taking over when one of us is missing.

37

Special thanks to Lorentz Engström, for the use of your office. It was essential to have aquiet place to work and to leave my messy paperwork.

Our good friends outside the lab: Aris, Katarina and Anna, Quny Wunny, Gao Ling andFei-Fei, Els and Cecilia for good diners and pleasant times together.

My parents, for love, support and encouragement.My sisters. Nannette and Christine.

Gijs, for your love, mental help, support and also for controlling and entertaining ourdaughter during the hectic times, keeping house, providing a computer at home andhelping me with the lay-out and many other things. Vera, for making me forget all about this.

38

REFERENCES

1. Bouck, N., Stellmach, V., and Hsu, S. C. How tumors become angiogenic. Adv Cancer Res, 69:135-174, 1996.

2. Hanahan, D. and Folkman, J. Patterns and emerging mechanisms of the angiogenic switch duringtumorigenesis. Cell, 86: 353-364., 1996.

3. Folkman, J. Tumor angiogenesis. In Cancer Medicine, p. pp. 41-64. Baltimore: MD: Williams andWilkins, 1997.

4. Hanahan, D. and Weinberg, R. A. The hallmarks of cancer. Cell, 100: 57-70., 2000.5. Bishop, J. M. a. W., R. A. Molecular Oncology. New York: Scientific American Inc., 1996.6. Renan, M. J. How many mutations are required for tumorigenesis? Implications from human cancer

data. Mol Carcinog, 7: 139-146, 1993.7. Venkatesh, Y. S., Ordonez, N. G., Schultz, P. N., Hickey, R. C., Goepfert, H., and Samaan, N. A.

Anaplastic carcinoma of the thyroid. A clinicopathologic study of 121 cases. Cancer, 66: 321-330.,1990.

8. Ain, K. B. Anaplastic thyroid carcinoma: behavior, biology, and therapeutic approaches. Thyroid, 8:715-726., 1998.

9. Demeter, J. G., De Jong, S. A., Lawrence, A. M., and Paloyan, E. Anaplastic thyroid carcinoma: riskfactors and outcome. Surgery, 110: 956-961; discussion 961-953., 1991.

10. LaFlamme, S. E. and Auer, K. L. Integrin signaling. Semin Cancer Biol, 7: 111-118., 1996.11. Weaver, V. M., Fischer, A. H., Peterson, O. W., and Bissell, M. J. The importance of the

microenvironment in breast cancer progression: recapitulation of mammary tumorigenesis using aunique human mammary epithelial cell model and a three-dimensional culture assay. Biochem CellBiol, 74: 833-851, 1996.

12. Frisch, S. M. and Francis, H. Disruption of epithelial cell-matrix interactions induces apoptosis. JCell Biol, 124: 619-626., 1994.

13. Folkman, J. Angiogenesis in cancer, vascular, rheumatoid and other disease. Nat Med, 1: 27-31.,1995.

14. Senger, D. R., Perruzzi, C. A., Feder, J., and Dvorak, H. F. A highly conserved vascularpermeability factor secreted by a variety of human and rodent tumor cell lines. Cancer Res, 46:5629-5632., 1986.

15. Kandel, J., Bossy-Wetzel, E., Radvanyi, F., Klagsbrun, M., Folkman, J., and Hanahan, D.Neovascularization is associated with a switch to the export of bFGF in the multistep developmentof fibrosarcoma. Cell, 66: 1095-1104., 1991.

16. Gregoire, M. and Lieubeau, B. The role of fibroblasts in tumor behavior. Cancer Metastasis Rev, 14:339-350., 1995.

17. Sappino, A. P., Schurch, W., and Gabbiani, G. Differentiation repertoire of fibroblastic cells:expression of cytoskeletal proteins as marker of phenotypic modulations. Lab Invest, 63: 144-161.,1990.

18. Nicolson, G. L. Cancer metastasis. Organ colonization and the cell-surface properties of malignantcells. Biochim Biophys Acta, 695: 113-176., 1982.

39

19. Schmitt, M., Janicke, F., Moniwa, N., Chucholowski, N., Pache, L., and Graeff, H. Tumor-associated urokinase-type plasminogen activator: biological and clinical significance. Biol ChemHoppe Seyler, 373: 611-622., 1992.

20. Clavel, C., Polette, M., Doco, M., Binninger, I., and Birembaut, P. Immunolocalization of matrixmetallo-proteinases and their tissue inhibitor in human mammary pathology. Bull Cancer, 79: 261-270, 1992.

21. Laiho, M., Saksela, O., Andreasen, P. A., and Keski-Oja, J. Enhanced production and extracellulardeposition of the endothelial- type plasminogen activator inhibitor in cultured human lungfibroblasts by transforming growth factor- β. J Cell Biol, 103: 2403-2410., 1986.

22. Pyke, C., Kristensen, P., Ralfkiaer, E., Eriksen, J., and Dano, K. The plasminogen activation systemin human colon cancer: messenger RNA for the inhibitor PAI-1 is located in endothelial cells in thetumor stroma. Cancer Res, 51: 4067-4071., 1991.

23. Reilly, C. F. and McFall, R. C. Platelet-derived growth factor and transforming growth factor- βregulate plasminogen activator inhibitor-1 synthesis in vascular smooth muscle cells. J Biol Chem,266: 9419-9427., 1991.

24. Ellerbroek, S. M. and Stack, M. S. Membrane associated matrix metalloproteinases in metastasis.Bioessays, 21: 940-949., 1999.

25. Curran, S. and Murray, G. I. Matrix metalloproteinases in tumour invasion and metastasis. J Pathol,189: 300-308., 1999.

26. Nagy, J. A., Brown, L. F., Senger, D. R., Lanir, N., Van de Water, L., Dvorak, A. M., and Dvorak,H. F. Pathogenesis of tumor stroma generation: a critical role for leaky blood vessels and fibrindeposition. Biochim Biophys Acta, 948: 305-326., 1989.

27. Dvorak, H. F., Nagy, J. A., Feng, D., Brown, L. F., and Dvorak, A. M. Vascular permeabilityfactor/vascular endothelial growth factor and the significance of microvascular hyperpermeability inangiogenesis. Curr Top Microbiol Immunol, 237: 97-132, 1999.

28. Feng, D., Nagy, J. A., Pyne, K., Hammel, I., Dvorak, H. F., and Dvorak, A. M. Pathways ofmacromolecular extravasation across microvascular endothelium in response to VPF/VEGF andother vasoactive mediators. Microcirculation, 6: 23-44., 1999.

29. McDonald, D. M., Thurston, G., and Baluk, P. Endothelial gaps as sites for plasma leakage ininflammation. Microcirculation, 6: 7-22., 1999.

30. Michel, C. C. and Neal, C. R. Openings through endothelial cells associated with increasedmicrovascular permeability. Microcirculation, 6: 45-54., 1999.

31. Brown, L. F., Detmar, M., Claffey, K., Nagy, J. A., Feng, D., Dvorak, A. M., and Dvorak, H. F.Vascular permeability factor/vascular endothelial growth factor: a multifunctional angiogeniccytokine. Exs, 79: 233-269, 1997.

32. Dano, K., Andreasen, P. A., Grondahl-Hansen, J., Kristensen, P., Nielsen, L. S., and Skriver, L.Plasminogen activators, tissue degradation, and cancer. Adv Cancer Res, 44: 139-266, 1985.

33. Dvorak, H. F., Harvey, V. S., and McDonagh, J. Quantitation of fibrinogen influx and fibrindeposition and turnover in line 1 and line 10 guinea pig carcinomas. Cancer Res, 44: 3348-3354.,1984.

34. Sporn, M. B. and Roberts, A. B. Autocrine growth factors and cancer. Nature, 313: 745-747., 1985.35. Heldin, C. H. a. B. W. Role of Platelet-Derived Growth Factor In Vivo., 2nd edition. edition, p. 249-

273. New York: Plenum, 1996.36. Clark, R. A. F. The molecular and cellular biology of wound repair., 2nd edition edition. New York:

Plenum Press, 1996.

40

37. Brown, E. J. Phagocytosis. Bioessays, 17: 109-117., 1995.38. Leibovich, S. J. and Ross, R. The role of the macrophage in wound repair. A study with

hydrocortisone and antimacrophage serum. Am J Pathol, 78: 71-100., 1975.39. Rappolee, D. A., Mark, D., Banda, M. J., and Werb, Z. Wound macrophages express TGF-α and

other growth factors in vivo: analysis by mRNA phenotyping. Science, 241: 708-712., 1988.40. Riches, D. W. H. Macrophage involvement in wound repair, remodeling and fibrosis. In: R. A. F.

Clark (ed.), The molecular and cellular biology of wound repair., 2nd ed. edition, pp. 95-141. NewYork: Plenum Press, 1996.

41. Ross, R., Masuda, J., Raines, E. W., Gown, A. M., Katsuda, S., Sasahara, M., Malden, L. T.,Masuko, H., and Sato, H. Localization of PDGF-B protein in macrophages in all phases ofatherogenesis. Science, 248: 1009-1012., 1990.

42. Hunt, T. K. Wound healing and wound infection: theory and surgical practice. New York: Appleton-Century-Crofts, 1980.

43. Mignatti, P., Rifkin, D. B., Welgus, H. G., Parks, W. C. Proteinases and tissue remodeling. In: R. A.F. Clark (ed.), The molecular and cellular biology of wound repair., 2nd ed. edition, pp. 427-474.New York: Plenum Press, 1996.

44. Vaalamo, M., Mattila, L., Johansson, N., Kariniemi, A. L., Karjalainen-Lindsberg, M. L., Kahari, V.M., and Saarialho-Kere, U. Distinct populations of stromal cells express collagenase-3 (MMP-13)and collagenase-1 (MMP-1) in chronic ulcers but not in normally healing wounds. J InvestDermatol, 109: 96-101., 1997.

45. Reuterdahl, C., Sundberg, C., Rubin, K., Funa, K., and Gerdin, B. Tissue localization of β receptorsfor platelet-derived growth factor and platelet-derived growth factor B chain during wound repair inhumans. J Clin Invest, 91: 2065-2075., 1993.

46. Greiling, D. and Clark, R. A. Fibronectin provides a conduit for fibroblast transmigration fromcollagenous stroma into fibrin clot provisional matrix. J Cell Sci, 110: 861-870., 1997.

47. Toole, B. P. Proteoglycans and hyaluronan in morphogenesis and differentiation. In: E. D. Hay(ed.), Cell biology of extracellular matrix, 2nd ed. edition, pp. 305-341. New York: Plenum Press,1991.

48. Welch, M. P., Odland, G. F., and Clark, R. A. Temporal relationships of F-actin bundle formation,collagen and fibronectin matrix assembly, and fibronectin receptor expression to wound contraction.J Cell Biol, 110: 133-145., 1990.

49. Ross, R., Raines, E. W., and Bowen-Pope, D. F. The biology of platelet-derived growth factor. Cell,46: 155-169., 1986.

50. Heldin, C. H. and Westermark, B. Mechanism of action and in vivo role of platelet-derived growthfactor. Physiol Rev, 79: 1283-1316., 1999.

51. Ostman, A. and Heldin, C. H. Involvement of platelet-derived growth factor in disease: developmentof specific antagonists. Adv Cancer Res, 80: 1-38, 2001.

52. Sporn, M. B., Roberts, A. B., Wakefield, L. M., and Assoian, R. K. Transforming growth factor- β:biological function and chemical structure. Science, 233: 532-534., 1986.

53. Sundberg, C., Branting, M., Gerdin, B., and Rubin, K. Tumor cell and connective tissue cellinteractions in human colorectal adenocarcinoma. Transfer of platelet-derived growth factor-AB/BBto stromal cells. Am J Pathol, 151: 479-492., 1997.

54. Brown, L. F., Van de Water, L., Harvey, V. S., and Dvorak, H. F. Fibrinogen influx andaccumulation of cross-linked fibrin in healing wounds and in tumor stroma. Am J Pathol, 130: 455-465., 1988.

41

55. Paku, S. and Paweletz, N. First steps of tumor-related angiogenesis. Lab Invest, 65: 334-346.,1991.

56. Bosman, F. T., de Bruine, A., Flohil, C., van der Wurff, A., ten Kate, J., and Dinjens, W. W.Epithelial-stromal interactions in colon cancer. Int J Dev Biol, 37: 203-211., 1993.

57. Dvorak, H. F., Nagy, J. A., Dvorak, J. T., and Dvorak, A. M. Identification and characterization ofthe blood vessels of solid tumors that are leaky to circulating macromolecules. Am J Pathol, 133:95-109., 1988.

58. Ivarsson, M., Sundberg, C., Farrokhnia, N., Pertoft, H., Rubin, K., and Gerdin, B. Recruitment oftype I collagen producing cells from the microvasculature in vitro. Exp Cell Res, 229: 336-349.,1996.

59. Sundberg, C., Ivarsson, M., Gerdin, B., and Rubin, K. Pericytes as collagen-producing cells inexcessive dermal scarring. Lab Invest, 74: 452-466., 1996.

60. Fosang, A. J. a. H., T. E. Matrix Proteoglycans. In: W. D. Comper (ed.), Extracellular matrix,Volume 2, Molecular Components and Interactions, Vol. volume 2, pp. 200-229. Reading, U.K.:Harwood Academic Publishers, 1996.

61. van der Rest, M. a. G., R. Collagen family of proteins. Faseb J, 5: 2814-2823, 1991.62. Timpl, R. Macromolecular organization of basement membranes. Curr Opin Cell Biol, 8: 618-624.,

1996.63. Hay, E. D. Cell biology of the Extracellular Matrix, 2nd edition. New York: Plenum Press, 1991.64. Barsky, S. H., Rao, C. N., Grotendorst, G. R., and Liotta, L. A. Increased content of Type V

Collagen in desmoplasia of human breast carcinoma. Am J Pathol, 108: 276-283., 1982.65. Dziadek, M. Role of laminin-nidogen complexes in basement membrane formation during

embryonic development. Experientia, 51: 901-913., 1995.66. Damjanov, I., Damjanov, N., Knowles, B. B., and Engvall, E. Origin of laminin in the extracellular

matrix of human tumor xenografts in nude mice. Virchows Arch B Cell Pathol Incl Mol Pathol, 49:45-52, 1985.

67. Cliff, W. J. Observations on healing tissue: A combined light and electron microscopy investigation.Phil Trans Soc Lond B, 246: 305-325, 1963.

68. Cliff, W. J. The extra-endothelial cells of the blood vessel wall. Blood Vessels Biol. Struct. Func., 6:68-72, 1976.

69. Mirra, S. S. and Miles, M. L. Unusual pericytic proliferation in a meningotheliomatous meningioma:an ultrastructural study. Am J Surg Pathol, 6: 573-580., 1982.

70. D'Amore, P. A. Capillary growth: a two-cell system. Semin Cancer Biol, 3: 49-56., 1992.71. Rhodin, J. A. Ultrastructure of mammalian venous capillaries, venules, and small collecting veins. J

Ultrastruct Res, 25: 452-500., 1968.72. Rhodin, J. A. G. Architecture of the vessel wall., Vol. part 2, chapter 1, p. 1-30. Baltimore: Waverly

Press, 1980.73. Sims, D. E. The pericyte - a review. Tissue cell, 18: 153-174, 1986.74. Kuper, H., Adami, H. O., and Trichopoulos, D. Infections as a major preventable cause of human

cancer. J Intern Med, 248: 171-183., 2000.75. Di Carlo, E., Forni, G., Lollini, P., Colombo, M. P., Modesti, A., and Musiani, P. The intriguing role

of polymorphonuclear neutrophils in antitumor reactions. Blood, 97: 339-345., 2001.76. Wahl, L. M. and Kleinman, H. K. Tumor-associated macrophages as targets for cancer therapy. J

Natl Cancer Inst, 90: 1583-1584., 1998.77. Pepper, M. S. Lymphangiogenesis and tumor metastasis: myth or reality? Clin Cancer Res, 7: 462-

468., 2001.

42

78. Leu, A. J., Berk, D. A., Lymboussaki, A., Alitalo, K., and Jain, R. K. Absence of functionallymphatics within a murine sarcoma: a molecular and functional evaluation. Cancer Res, 60: 4324-4327., 2000.

79. Hartveit, E. Attenuated cells in breast stroma: the missing lymphatic system of the breast.Histopathology, 16: 533-543., 1990.

80. Beck, K. and Brodsky, B. Supercoiled protein motifs: the collagen triple-helix and the α- helicalcoiled coil. J Struct Biol, 122: 17-29, 1998.

81. Kramer, R. Z., Bella, J., Mayville, P., Brodsky, B., and Berman, H. M. Sequence dependentconformational variations of collagen triple-helical structure. Nat Struct Biol, 6: 454-457., 1999.

82. Vogel, K. G., Paulsson, M., and Heinegard, D. Specific inhibition of type I and type II collagenfibrillogenesis by the small proteoglycan of tendon. Biochem J, 223: 587-597., 1984.

83. Hedbom, E. and Heinegard, D. Interaction of a 59-kDa connective tissue matrix protein withcollagen I and collagen II. J Biol Chem, 264: 6898-6905., 1989.

84. Rada, J. A., Cornuet, P. K., and Hassell, J. R. Regulation of corneal collagen fibrillogenesis in vitroby corneal proteoglycan (lumican and decorin) core proteins. Exp Eye Res, 56: 635-648., 1993.

85. Vogel, K. G. and Trotter, J. A. The effect of proteoglycans on the morphology of collagen fibrilsformed in vitro. Coll Relat Res, 7: 105-114., 1987.

86. Kyriakides, T. R., Zhu, Y. H., Smith, L. T., Bain, S. D., Yang, Z., Lin, M. T., Danielson, K. G.,Iozzo, R. V., LaMarca, M., McKinney, C. E., Ginns, E. I., and Bornstein, P. Mice that lackthrombospondin 2 display connective tissue abnormalities that are associated with disorderedcollagen fibrillogenesis, an increased vascular density, and a bleeding diathesis. J Cell Biol, 140:419-430., 1998.

87. Kivirikko, K. I. Collagens and their abnormalities in a wide spectrum of diseases. Ann Med, 25:113-126., 1993.

88. Brown, J. C. and Timpl, R. The collagen superfamily. Int Arch Allergy Immunol, 107: 484-490.,1995.

89. Kadler, K. Extracellular matrix 1: Fibril-forming collagens. Protein Profile, 2: 491-619, 1995.90. Olsen, B. R. New insights into the function of collagens from genetic analysis. Curr Opin Cell Biol,

7: 720-727., 1995.91. Prockop, D. J. and Kivirikko, K. I. Collagens: molecular biology, diseases, and potentials for

therapy. Annu Rev Biochem, 64: 403-434, 1995.92. Bateman, J. F., Lamandé, S.R., Ramshaw, J.A.M. Collagen superfamily., Vol. volume 2, p. pp. 22-

67. Amsterdam: Harwood, 1996.93. McLaughlin, S. H. and Bulleid, N. J. Molecular recognition in procollagen chain assembly. Matrix

Biol, 16: 369-377., 1998.94. Bonfanti, L., Mironov, A. A., Jr., Martinez-Menarguez, J. A., Martella, O., Fusella, A., Baldassarre,

M., Buccione, R., Geuze, H. J., Mironov, A. A., and Luini, A. Procollagen traverses the Golgi stackwithout leaving the lumen of cisternae: evidence for cisternal maturation. Cell, 95: 993-1003., 1998.

95. Prockop, D. J., Sieron, A. L., and Li, S. W. Procollagen N-proteinase and procollagen C-proteinase.Two unusual metalloproteinases that are essential for procollagen processing probably haveimportant roles in development and cell signaling. Matrix Biol, 16: 399-408., 1998.

96. Kadler, K. E., Holmes, D. F., Trotter, J. A., and Chapman, J. A. Collagen fibril formation. BiochemJ, 316: 1-11., 1996.

43

97. Graham, H. K., Holmes, D. F., Watson, R. B., and Kadler, K. E. Identification of collagen fibrilfusion during vertebrate tendon morphogenesis. The process relies on unipolar fibrils and isregulated by collagen-proteoglycan interaction. J Mol Biol, 295: 891-902., 2000.

98. Smith-Mungo, L. I. and Kagan, H. M. Lysyl oxidase: properties, regulation and multiple functionsin biology. Matrix Biol, 16: 387-398., 1998.

99. Lamande, S. R. and Bateman, J. F. Procollagen folding and assembly: the role of endoplasmicreticulum enzymes and molecular chaperones. Semin Cell Dev Biol, 10: 455-464., 1999.

100. Nagata, K. Expression and function of heat shock protein 47: a collagen-specific molecularchaperone in the endoplasmic reticulum. Matrix Biol, 16: 379-386., 1998.

101. Tasab, M., Batten, M. R., and Bulleid, N. J. Hsp47: a molecular chaperone that interacts with andstabilizes correctly-folded procollagen. Embo J, 19: 2204-2211., 2000.

102. Nagai, N., Hosokawa, M., Itohara, S., Adachi, E., Matsushita, T., Hosokawa, N., and Nagata, K.Embryonic lethality of molecular chaperone hsp47 knockout mice is associated with defects incollagen biosynthesis. J Cell Biol, 150: 1499-1506., 2000.

103. Rossi, P., Karsenty, G., Roberts, A. B., Roche, N. S., Sporn, M. B., and de Crombrugghe, B. Anuclear factor 1 binding site mediates the transcriptional activation of a type I collagen promoter bytransforming growth factor- β. Cell, 52: 405-414., 1988.

104. Inagaki, Y., Truter, S., and Ramirez, F. Transforming growth factor- β stimulates α 2(I) collagengene expression through a cis-acting element that contains an Sp1-binding site. J Biol Chem, 269:14828-14834., 1994.

105. Jimenez, S. A., Varga, J., Olsen, A., Li, L., Diaz, A., Herhal, J., and Koch, J. Functional analysis ofhuman α 1(I) procollagen gene promoter. Differential activity in collagen-producing and -nonproducing cells and response to transforming growth factor β 1. J Biol Chem, 269: 12684-12691., 1994.

106. Wrana, J. L., Overall, C. M., and Sodek, J. Regulation of the expression of a secreted acidic proteinrich in cysteine (SPARC) in human fibroblasts by transforming growth factor β. Comparison oftranscriptional and post-transcriptional control with fibronectin and type I collagen. Eur J Biochem,197: 519-528., 1991.

107. Armendariz-Borunda, J., Katayama, K., and Seyer, J. M. Transcriptional mechanisms of type Icollagen gene expression are differentially regulated by interleukin-1 β, tumor necrosis factor α,and transforming growth factor β in Ito cells. J Biol Chem, 267: 14316-14321., 1992.

108. Fine, A. and Goldstein, R. H. Regulation of type I collagen mRNA translation by TGF- β. RegImmunol, 5: 218-224., 1993.

109. Lewinski, A., Pawlikowski, M., and Cardinali, D. P. Thyroid growth-stimulating and growth-inhibiting factors. Biol Signals, 2: 313-351., 1993.

110. Shore, R. E. Issues and epidemiological evidence regarding radiation-induced thyroid cancer. RadiatRes, 131: 98-111., 1992.

111. Bi, J. and Lu, B. Advances in the diagnosis and management of thyroid neoplasms. Curr OpinOncol, 12: 54-59., 2000.

112. Ljungberg, O. Biopsy pathology of the thyroid and parathyroid: Chapman & Hall Medical., 1992.113. Tallini, G., Santoro, M., Helie, M., Carlomagno, F., Salvatore, G., Chiappetta, G., Carcangiu, M. L.,

and Fusco, A. RET/PTC oncogene activation defines a subset of papillary thyroid carcinomaslacking evidence of progression to poorly differentiated or undifferentiated tumor phenotypes. ClinCancer Res, 4: 287-294., 1998.

44

114. Basolo, F., Molinaro, E., Agate, L., Pinchera, A., Pollina, L., Chiappetta, G., Monaco, C.,Santoro, M., Fusco, A., Miccoli, P., Elisei, R., Capezzone, M., and Pacini, F. RET proteinexpression has no prognostic impact on the long-term outcome of papillary thyroid carcinoma. Eur JEndocrinol, 145: 599-604., 2001.

115. Mizukami, Y., Michigishi, T., Nonomura, A., Noguchi, M., and Nakamura, H. Thyroid carcinoma:clinical, pathologic correlations. Crit Rev Oncol Hematol, 18: 67-102., 1995.

116. De Groot, L. J., Reed Larsen, P. and Henneman, G. The thyroid and its diseases, 6 ed. edition. NewYork: Churchill Livingstone, 1996.

117. Nishida, T., Katayama, S., Tsujimoto, M., Nakamura, J., and Matsuda, H. Clinicopathologicalsignificance of poorly differentiated thyroid carcinoma. Am J Surg Pathol, 23: 205-211., 1999.

118. Fonseca, E., Soares, P., Rossi, S., and Sobrinho-Simoes, M. Prognostic factors in thyroidcarcinomas. Verh Dtsch Ges Pathol, 81: 82-96, 1997.

119. Carcangiu, M. L., Zampi, G., Pupi, A., Castagnoli, A., and Rosai, J. Papillary carcinoma of thethyroid. A clinicopathologic study of 241 cases treated at the University of Florence, Italy. Cancer,55: 805-828., 1985.

120. Massague, J., Blain, S. W., and Lo, R. S. TGF β signaling in growth control, cancer, and heritabledisorders. Cell, 103: 295-309., 2000.

121. Roberts, A. B. and Sporn, M. B. Differential expression of the TGF- β isoforms in embryogenesissuggests specific roles in developing and adult tissues. Mol Reprod Dev, 32: 91-98., 1992.

122. Piek, E., Heldin, C. H., and Ten Dijke, P. Specificity, diversity, and regulation in TGF- βsuperfamily signaling. Faseb J, 13: 2105-2124., 1999.

123. Munger, J. S., Harpel, J. G., Gleizes, P. E., Mazzieri, R., Nunes, I., and Rifkin, D. B. Latenttransforming growth factor- β: structural features and mechanisms of activation. Kidney Int, 51:1376-1382., 1997.

124. Munger, J. S., Huang, X., Kawakatsu, H., Griffiths, M. J., Dalton, S. L., Wu, J., Pittet, J. F.,Kaminski, N., Garat, C., Matthay, M. A., Rifkin, D. B., and Sheppard, D. The integrin α v β 6 bindsand activates latent TGF β 1: a mechanism for regulating pulmonary inflammation and fibrosis.Cell, 96: 319-328., 1999.

125. Crawford, S. E., Stellmach, V., Murphy-Ullrich, J. E., Ribeiro, S. M., Lawler, J., Hynes, R. O.,Boivin, G. P., and Bouck, N. Thrombospondin-1 is a major activator of TGF- β1 in vivo. Cell, 93:1159-1170., 1998.

126. Wrana, J. L., Attisano, L., Wieser, R., Ventura, F., and Massague, J. Mechanism of activation of theTGF- β receptor. Nature, 370: 341-347., 1994.

127. Lopez-Casillas, F., Wrana, J. L., and Massague, J. βglycan presents ligand to the TGF β signalingreceptor. Cell, 73: 1435-1444., 1993.

128. Massague, J. TGF- β signal transduction. Annu Rev Biochem, 67: 753-791, 1998.129. Heldin, C. H., Miyazono, K., and ten Dijke, P. TGF- β signalling from cell membrane to nucleus

through SMAD proteins. Nature, 390: 465-471., 1997.130. Nakao, A., Imamura, T., Souchelnytskyi, S., Kawabata, M., Ishisaki, A., Oeda, E., Tamaki, K.,

Hanai, J., Heldin, C. H., Miyazono, K., and ten Dijke, P. TGF- β receptor-mediated signallingthrough Smad2, Smad3 and Smad4. Embo J, 16: 5353-5362., 1997.

131. Macias-Silva, M., Abdollah, S., Hoodless, P. A., Pirone, R., Attisano, L., and Wrana, J. L. MADR2is a substrate of the TGF β receptor and its phosphorylation is required for nuclear accumulation andsignaling. Cell, 87: 1215-1224., 1996.

45

132. Nakao, A., Afrakhte, M., Moren, A., Nakayama, T., Christian, J. L., Heuchel, R., Itoh, S.,Kawabata, M., Heldin, N. E., Heldin, C. H., and ten Dijke, P. Identification of Smad7, a TGF β-inducible antagonist of TGF- β signalling. Nature, 389: 631-635., 1997.

133. Hayashi, H., Abdollah, S., Qiu, Y., Cai, J., Xu, Y. Y., Grinnell, B. W., Richardson, M. A., Topper, J.N., Gimbrone, M. A., Jr., Wrana, J. L., and Falb, D. The MAD-related protein Smad7 associateswith the TGF β receptor and functions as an antagonist of TGF β signaling. Cell, 89: 1165-1173.,1997.

134. Chen, Y. G., Liu, F., and Massague, J. Mechanism of TGF β receptor inhibition by FKBP12. EmboJ, 16: 3866-3876., 1997.

135. Onichtchouk, D., Chen, Y. G., Dosch, R., Gawantka, V., Delius, H., Massague, J., and Niehrs, C.Silencing of TGF- β signalling by the pseudoreceptor BAMBI. Nature, 401: 480-485., 1999.

136. Datta, P. K. and Moses, H. L. STRAP and Smad7 synergize in the inhibition of transforming growthfactor β signaling. Mol Cell Biol, 20: 3157-3167., 2000.

137. Charng, M. J., Zhang, D., Kinnunen, P., and Schneider, M. D. A novel protein distinguishesbetween quiescent and activated forms of the type I transforming growth factor β receptor. J BiolChem, 273: 9365-9368., 1998.

138. Tsukazaki, T., Chiang, T. A., Davison, A. F., Attisano, L., and Wrana, J. L. SARA, a FYVE domainprotein that recruits Smad2 to the TGF β receptor. Cell, 95: 779-791., 1998.

139. Satterwhite, D. J. and Moses, H. L. Mechanisms of transforming growth factor- β 1-induced cellcycle arrest. Invasion Metastasis, 14: 309-318, 1994.

140. Massague, J., Cheifetz, S., Laiho, M., Ralph, D. A., Weis, F. M., and Zentella, A. Transforminggrowth factor- β. Cancer Surv, 12: 81-103, 1992.

141. Heldin, N. E., Bergstrom, D., Hermansson, A., Bergenstrahle, A., Nakao, A., Westermark, B., andten Dijke, P. Lack of responsiveness to TGF- β1 in a thyroid carcinoma cell line with functionaltype I and type II TGF- β receptors and Smad proteins, suggests a novel mechanism for TGF- βinsensitivity in carcinoma cells. Mol Cell Endocrinol, 153: 79-90., 1999.

142. Franzen, A., Piek, E., Westermark, B., ten Dijke, P., and Heldin, N. E. Expression of transforminggrowth factor- β1, activin A, and their receptors in thyroid follicle cells: negative regulation ofthyrocyte growth and function. Endocrinology, 140: 4300-4310., 1999.

143. Battegay, E. J., Raines, E. W., Seifert, R. A., Bowen-Pope, D. F., and Ross, R. TGF- β inducesbimodal proliferation of connective tissue cells via complex control of an autocrine PDGF loop.Cell, 63: 515-524., 1990.

144. Weinberg, R. A. The retinoblastoma protein and cell cycle control. Cell, 81: 323-330., 1995.145. Koff, A., Ohtsuki, M., Polyak, K., Roberts, J. M., and Massague, J. Negative regulation of G1 in

mammalian cells: inhibition of cyclin E- dependent kinase by TGF- β. Science, 260: 536-539.,1993.

146. Reynisdottir, I., Polyak, K., Iavarone, A., and Massague, J. Kip/Cip and Ink4 Cdk inhibitorscooperate to induce cell cycle arrest in response to TGF- β. Genes Dev, 9: 1831-1845., 1995.

147. Datto, M. B., Yu, Y., and Wang, X. F. Functional analysis of the transforming growth factor βresponsive elements in the WAF1/Cip1/p21 promoter. J Biol Chem, 270: 28623-28628., 1995.

148. Bishayee, S., Majumdar, S., Khire, J., and Das, M. Ligand-induced dimerization of the platelet-derived growth factor receptor. Monomer-dimer interconversion occurs independent of receptorphosphorylation. J Biol Chem, 264: 11699-11705., 1989.

46

149. Heldin, C. H., Ernlund, A., Rorsman, C., and Ronnstrand, L. Dimerization of B-type platelet-derived growth factor receptors occurs after ligand binding and is closely associated with receptorkinase activation. J Biol Chem, 264: 8905-8912., 1989.

150. Seifert, R. A., Hart, C. E., Phillips, P. E., Forstrom, J. W., Ross, R., Murray, M. J., and Bowen-Pope,D. F. Two different subunits associate to create isoform-specific platelet- derived growth factorreceptors. J Biol Chem, 264: 8771-8778., 1989.

151. Betsholtz, C., Johnsson, A., Heldin, C. H., Westermark, B., Lind, P., Urdea, M. S., Eddy, R., Shows,T. B., Philpott, K., Mellor, A. L., and et al. cDNA sequence and chromosomal localization of humanplatelet-derived growth factor A-chain and its expression in tumour cell lines. Nature, 320: 695-699., 1986.

152. Swan, D. C., McBride, O. W., Robbins, K. C., Keithley, D. A., Reddy, E. P., and Aaronson, S. A.Chromosomal mapping of the simian sarcoma virus onc gene analogue in human cells. Proc NatlAcad Sci U S A, 79: 4691-4695., 1982.

153. Li, X., Ponten, A., Aase, K., Karlsson, L., Abramsson, A., Uutela, M., Backstrom, G., Hellstrom,M., Bostrom, H., Li, H., Soriano, P., Betsholtz, C., Heldin, C. H., Alitalo, K., Ostman, A., andEriksson, U. PDGF-C is a new protease-activated ligand for the PDGF α-receptor. Nat Cell Biol, 2:302-309., 2000.

154. LaRochelle, W. J., Jeffers, M., McDonald, W. F., Chillakuru, R. A., Giese, N. A., Lokker, N. A.,Sullivan, C., Boldog, F. L., Yang, M., Vernet, C., Burgess, C. E., Fernandes, E., Deegler, L. L.,Rittman, B., Shimkets, J., Shimkets, R. A., Rothberg, J. M., and Lichenstein, H. S. PDGF-D, a newprotease-activated growth factor. Nat Cell Biol, 3: 517-521., 2001.

155. Bergsten, E., Uutela, M., Li, X., Pietras, K., Ostman, A., Heldin, C. H., Alitalo, K., and Eriksson, U.PDGF-D is a specific, protease-activated ligand for the PDGF β- receptor. Nat Cell Biol, 3: 512-516., 2001.

156. Kazlauskas, A. Receptor tyrosine kinases and their targets. Curr Opin Genet Dev, 4: 5-14., 1994.157. Claesson-Welsh, L. Platelet-derived growth factor receptor signals. J Biol Chem, 269: 32023-

32026., 1994.158. Wells, R. G. Fibrogenesis. V. TGF- β signaling pathways. Am J Physiol Gastrointest Liver Physiol,

279: G845-850., 2000.159. Vignaud, J. M., Allam, M., Martinet, N., Pech, M., Plenat, F., and Martinet, Y. Presence of platelet-

derived growth factor in normal and fibrotic lung is specifically associated with interstitialmacrophages, while both interstitial macrophages and alveolar epithelial cells express the c- sisproto-oncogene. Am J Respir Cell Mol Biol, 5: 531-538., 1991.

160. Branton, M. H. and Kopp, J. B. TGF- β and fibrosis. Microbes Infect, 1: 1349-1365., 1999.161. Isaka, Y., Fujiwara, Y., Ueda, N., Kaneda, Y., Kamada, T., and Imai, E. Glomerulosclerosis induced

by in vivo transfection of transforming growth factor- β or platelet-derived growth factor gene intothe rat kidney. J Clin Invest, 92: 2597-2601., 1993.

162. Schmid, P., Kunz, S., Cerletti, N., McMaster, G., and Cox, D. Injury induced expression of TGF- β1 mRNA is enhanced by exogenously applied TGF- β S. Biochem Biophys Res Commun, 194: 399-406., 1993.

163. Edwards, D. R., Murphy, G., Reynolds, J. J., Whitham, S. E., Docherty, A. J., Angel, P., and Heath,J. K. Transforming growth factor β modulates the expression of collagenase and metalloproteinaseinhibitor. Embo J, 6: 1899-1904., 1987.

47

164. Overall, C. M., Wrana, J. L., and Sodek, J. Independent regulation of collagenase, 72-kDaprogelatinase, and metalloendoproteinase inhibitor expression in human fibroblasts by transforminggrowth factor- β. J Biol Chem, 264: 1860-1869., 1989.

165. Assoian, R. K. and Sporn, M. B. Type β transforming growth factor in human platelets: releaseduring platelet degranulation and action on vascular smooth muscle cells. J Cell Biol, 102: 1217-1223., 1986.

166. Antoniades, H. N., Scher, C. D., and Stiles, C. D. Purification of human platelet-derived growthfactor. Proc Natl Acad Sci U S A, 76: 1809-1813., 1979.

167. Assoian, R. K., Fleurdelys, B. E., Stevenson, H. C., Miller, P. J., Madtes, D. K., Raines, E. W.,Ross, R., and Sporn, M. B. Expression and secretion of type β transforming growth factor byactivated human macrophages. Proc Natl Acad Sci U S A, 84: 6020-6024., 1987.

168. Kehrl, J. H., Wakefield, L. M., Roberts, A. B., Jakowlew, S., Alvarez-Mon, M., Derynck, R., Sporn,M. B., and Fauci, A. S. Production of transforming growth factor β by human T lymphocytes andits potential role in the regulation of T cell growth. J Exp Med, 163: 1037-1050., 1986.

169. Wahl, S. M., Hunt, D. A., Wakefield, L. M., McCartney-Francis, N., Wahl, L. M., Roberts, A. B.,and Sporn, M. B. Transforming growth factor type β induces monocyte chemotaxis and growthfactor production. Proc Natl Acad Sci U S A, 84: 5788-5792., 1987.

170. Postlethwaite, A. E., Keski-Oja, J., Moses, H. L., and Kang, A. H. Stimulation of the chemotacticmigration of human fibroblasts by transforming growth factor β. J Exp Med, 165: 251-256., 1987.

171. Bauer, E. A., Cooper, T. W., Huang, J. S., Altman, J., and Deuel, T. F. Stimulation of in vitro humanskin collagenase expression by platelet- derived growth factor. Proc Natl Acad Sci U S A, 82: 4132-4136., 1985.

172. Gullberg, D., Tingstrom, A., Thuresson, A. C., Olsson, L., Terracio, L., Borg, T. K., and Rubin, K.β 1 integrin-mediated collagen gel contraction is stimulated by PDGF. Exp Cell Res, 186: 264-272.,1990.

173. Mustoe, T. A., Pierce, G. F., Morishima, C., and Deuel, T. F. Growth factor-induced acceleration oftissue repair through direct and inductive activities in a rabbit dermal ulcer model. J Clin Invest, 87:694-703., 1991.

174. Pierce, G. F., Vande Berg, J., Rudolph, R., Tarpley, J., and Mustoe, T. A. Platelet-derived growthfactor-BB and transforming growth factor β 1 selectively modulate glycosaminoglycans, collagen,and myofibroblasts in excisional wounds. Am J Pathol, 138: 629-646., 1991.

175. Pierce, G. F., Brown, D., and Mustoe, T. A. Quantitative analysis of inflammatory cell influx,procollagen type I synthesis, and collagen cross-linking in incisional wounds: influence of PDGF-BB and TGF- β 1 therapy. J Lab Clin Med, 117: 373-382., 1991.

176. Hermansson, M., Nister, M., Betsholtz, C., Heldin, C. H., Westermark, B., and Funa, K. Endothelialcell hyperplasia in human glioblastoma: coexpression of mRNA for platelet-derived growth factor(PDGF) B chain and PDGF receptor suggests autocrine growth stimulation. Proc Natl Acad Sci U SA, 85: 7748-7752., 1988.

177. Hermanson, M., Funa, K., Hartman, M., Claesson-Welsh, L., Heldin, C. H., Westermark, B., andNister, M. Platelet-derived growth factor and its receptors in human glioma tissue: expression ofmessenger RNA and protein suggests the presence of autocrine and paracrine loops. Cancer Res, 52:3213-3219., 1992.

178. Plate, K. H., Breier, G., Farrell, C. L., and Risau, W. Platelet-derived growth factor receptor- β isinduced during tumor development and upregulated during tumor progression in endothelial cells inhuman gliomas. Lab Invest, 67: 529-534., 1992.

48

179.Westermark, B., Heldin, C. H., and Nister, M. Platelet-derived growth factor in human glioma.Glia, 15: 257-263., 1995.

180. Forsberg, K., Valyi-Nagy, I., Heldin, C. H., Herlyn, M., and Westermark, B. Platelet-derived growthfactor (PDGF) in oncogenesis: development of a vascular connective tissue stroma inxenotransplanted human melanoma producing PDGF-BB. Proc Natl Acad Sci U S A, 90: 393-397.,1993.

181. Fleming, T. P., Saxena, A., Clark, W. C., Robertson, J. T., Oldfield, E. H., Aaronson, S. A., and Ali,I. U. Amplification and/or overexpression of platelet-derived growth factor receptors and epidermalgrowth factor receptor in human glial tumors. Cancer Res, 52: 4550-4553., 1992.

182. Hermanson, M., Funa, K., Koopmann, J., Maintz, D., Waha, A., Westermark, B., Heldin, C. H.,Wiestler, O. D., Louis, D. N., von Deimling, A., and Nister, M. Association of loss ofheterozygosity on chromosome 17p with high platelet-derived growth factor α receptor expressionin human malignant gliomas. Cancer Res, 56: 164-171., 1996.

183. Kumabe, T., Sohma, Y., Kayama, T., Yoshimoto, T., and Yamamoto, T. Amplification of α-platelet-derived growth factor receptor gene lacking an exon coding for a portion of the extracellularregion in a primary brain tumor of glial origin. Oncogene, 7: 627-633., 1992.

184. Simon, M. P., Pedeutour, F., Sirvent, N., Grosgeorge, J., Minoletti, F., Coindre, J. M., Terrier-Lacombe, M. J., Mandahl, N., Craver, R. D., Blin, N., Sozzi, G., Turc-Carel, C., O'Brien, K. P.,Kedra, D., Fransson, I., Guilbaud, C., and Dumanski, J. P. Deregulation of the platelet-derivedgrowth factor B-chain gene via fusion with collagen gene COL1A1 in dermatofibrosarcomaprotuberans and giant-cell fibroblastoma. Nat Genet, 15: 95-98., 1997.

185. Skobe, M. and Fusenig, N. E. Tumorigenic conversion of immortal human keratinocytes throughstromal cell activation. Proc Natl Acad Sci U S A, 95: 1050-1055., 1998.

186. Tokunaga, H., Lee, D. H., Kim, I. Y., Wheeler, T. M., and Lerner, S. P. Decreased expression oftransforming growth factor β receptor type I is associated with poor prognosis in bladdertransitional cell carcinoma patients. Clin Cancer Res, 5: 2520-2525., 1999.

187. Miyaki, M., Iijima, T., Konishi, M., Sakai, K., Ishii, A., Yasuno, M., Hishima, T., Koike, M.,Shitara, N., Iwama, T., Utsunomiya, J., Kuroki, T., and Mori, T. Higher frequency of Smad4 genemutation in human colorectal cancer with distant metastasis. Oncogene, 18: 3098-3103., 1999.

188. Wieser, R. The transforming growth factor- β signaling pathway in tumorigenesis. Curr Opin Oncol,13: 70-77., 2001.

189. Tang, B., Bottinger, E. P., Jakowlew, S. B., Bagnall, K. M., Mariano, J., Anver, M. R., Letterio, J.J., and Wakefield, L. M. Transforming growth factor- β1 is a new form of tumor suppressor withtrue haploid insufficiency. Nat Med, 4: 802-807., 1998.

190. Gold, L. I. The role for Transforming Growth Factor-β (TGF-β) in human cancer. Crit Rev Oncog,10: 303-360, 1999.

191. Park, J. A., Wang, E., Kurt, R. A., Schluter, S. F., Hersh, E. M., and Akporiaye, E. T. Expression ofan antisense transforming growth factor- β1 transgene reduces tumorigenicity of EMT6 mammarytumor cells. Cancer Gene Ther, 4: 42-50., 1997.

192. Cui, W., Fowlis, D. J., Bryson, S., Duffie, E., Ireland, H., Balmain, A., and Akhurst, R. J. TGF β1inhibits the formation of benign skin tumors, but enhances progression to invasive spindlecarcinomas in transgenic mice. Cell, 86: 531-542., 1996.

193. Oft, M., Heider, K. H., and Beug, H. TGF β signaling is necessary for carcinoma cell invasivenessand metastasis. Curr Biol, 8: 1243-1252., 1998.

49

194.Yin, J. J., Selander, K., Chirgwin, J. M., Dallas, M., Grubbs, B. G., Wieser, R., Massague, J.,Mundy, G. R., and Guise, T. A. TGF- β signaling blockade inhibits PTHrP secretion by breastcancer cells and bone metastases development. J Clin Invest, 103: 197-206., 1999.

195. Aukland, K. and Reed, R. K. Interstitial-lymphatic mechanisms in the control of extracellular fluidvolume. Physiol Rev, 73: 1-78, 1993.

196. Reed, R. K., Rubin, K., Wiig, H., and Rodt, S. A. Blockade of β 1-integrins in skin causes edemathrough lowering of interstitial fluid pressure. Circ Res, 71: 978-983., 1992.

197. Rodt, S. A., Ahlen, K., Berg, A., Rubin, K., and Reed, R. K. A novel physiological function forplatelet-derived growth factor-BB in rat dermis. J Physiol, 495: 193-200., 1996.

198. Heuchel, R., Berg, A., Tallquist, M., Ahlen, K., Reed, R. K., Rubin, K., Claesson-Welsh, L., Heldin,C. H., and Soriano, P. Platelet-derived growth factor β receptor regulates interstitial fluidhomeostasis through phosphatidylinositol-3' kinase signaling. Proc Natl Acad Sci U S A, 96: 11410-11415., 1999.

199. Berg, A., Ekwall, A. K., Rubin, K., Stjernschantz, J., and Reed, R. K. Effect of PGE1, PGI2, andPGF2 α analogs on collagen gel compaction in vitro and interstitial pressure in vivo. Am J Physiol,274: H663-671, 1998.

200. Bell, E., Ivarsson, B., and Merrill, C. Production of a tissue-like structure by contraction of collagenlattices by human fibroblasts of different proliferative potential in vitro. Proc Natl Acad Sci U S A,76: 1274-1278., 1979.

201. Guidry, C. and Grinnell, F. Studies on the mechanism of hydrated collagen gel reorganization byhuman skin fibroblasts. J Cell Sci, 79: 67-81., 1985.

202. Rodt, S. A., Reed, R. K., Ljungström, M., Gustafsson, T. O., and Rubin, K. The anti-inflammatoryagent α-trinositol exerts its edema- preventing effects through modulation of β1 integrin function.Circ Res, 75: 942-948, 1994.

203. Åhlen, K., Berg, A., Stiger, F., Tengholm, A., Siegbahn, A., Gylfe, E., Reed, R. K., and Rubin, K.Cell interactions with collagen matrices in vivo and in vitro depend on phosphatidylinositol 3-kinaseand free cytoplasmic calcium. Cell Adhes Commun, 5: 461-473, 1998.

204. Rubin, K., Gullberg, D., Tomasini-Johansson, B., Reed, R. K., Rydén, C. and Borg, T. K. Molecularrecognition of the extracellular matrix by cell surface receptors. In: W. D. Comper (ed.),Extracellular matrix, Volume 2, Molecular Components and Interactions., Vol. volume 2, pp. 262-309. Reading, U.K.: Harwood Academic Publishers, 1996.

205. Reed, R. K., Woie, K., and Rubin, K. Integrins and control of interstitial fluid pressure. NewsPhysiol. Sci., 12: 42-48, 1997.

206. Meyer, F. A. Macromolecular basis of globular protein exclusion and of swelling pressure in looseconnective tissue (umbilical cord). Biochim Biophys Acta, 755: 388-399., 1983.

207. Gutmann, R., Leunig, M., Feyh, J., Goetz, A. E., Messmer, K., Kastenbauer, E., and Jain, R. K.Interstitial hypertension in head and neck tumors in patients: correlation with tumor size. CancerRes, 52: 1993-1995., 1992.

208. Less, J. R., Posner, M. C., Boucher, Y., Borochovitz, D., Wolmark, N., and Jain, R. K. Interstitialhypertension in human breast and colorectal tumors. Cancer Res, 52: 6371-6374., 1992.

209. Nathanson, S. D. and Nelson, L. Interstitial fluid pressure in breast cancer, benign breast conditions,and breast parenchyma. Ann Surg Oncol, 1: 333-338., 1994.

210. Jain, R. K. Delivery of molecular medicine to solid tumors. Science, 271: 1079-1080, 1996.

50

211.Netti, P. A., Hamberg, L. M., Babich, J. W., Kierstead, D., Graham, W., Hunter, G. J., Wolf, G.L., Fischman, A., Boucher, Y., and Jain, R. K. Enhancement of fluid filtration across tumor vessels:implication for delivery of macromolecules. Proc Natl Acad Sci U S A, 96: 3137-3142., 1999.

212. Baxter, L. T. and Jain, R. K. Transport of fluid and macromolecules in tumors. I. Role of interstitialpressure and convection. Microvasc Res, 37: 77-104., 1989.

213. Carmeliet, P. and Jain, R. K. Angiogenesis in cancer and other diseases. Nature, 407: 249-257.,2000.

214. Griffon-Etienne, G., Boucher, Y., Brekken, C., Suit, H. D., and Jain, R. K. Taxane-inducedapoptosis decompresses blood vessels and lowers interstitial fluid pressure in solid tumors: clinicalimplications. Cancer Res, 59: 3776-3782., 1999.

215. Netti, P. A., Berk, D. A., Swartz, M. A., Grodzinsky, A. J., and Jain, R. K. Role of extracellularmatrix assembly in interstitial transport in solid tumors. Cancer Res, 60: 2497-2503., 2000.

216. Lee, I., Boucher, Y., and Jain, R. K. Nicotinamide can lower tumor interstitial fluid pressure:mechanistic and therapeutic implications. Cancer Res, 52: 3237-3240., 1992.

217. Kristensen, C. A., Nozue, M., Boucher, Y., and Jain, R. K. Reduction of interstitial fluid pressureafter TNF-α treatment of three human melanoma xenografts. Br J Cancer, 74: 533-536., 1996.

218. Kristjansen, P. E., Boucher, Y., and Jain, R. K. Dexamethasone reduces the interstitial fluid pressurein a human colon adenocarcinoma xenograft. Cancer Res, 53: 4764-4766., 1993.

219. Rubin, K., Sjoquist, M., Gustafsson, A. M., Isaksson, B., Salvessen, G., and Reed, R. K. Loweringof tumoral interstitial fluid pressure by prostaglandin E(1) is paralleled by an increased uptake of(51)Cr-EDTA. Int J Cancer, 86: 636-643., 2000.

220. Pietras, K., Ostman, A., Sjoquist, M., Buchdunger, E., Reed, R. K., Heldin, C. H., and Rubin, K.Inhibition of platelet-derived growth factor receptors reduces interstitial hypertension and increasestranscapillary transport in tumors. Cancer Res, 61: 2929-2934., 2001.

221. Folkman, J. and Moscona, A. Role of cell shape in growth control. Nature, 273: 345-349., 1978.222. Ivarsson, M., McWhirter, A., Borg, T. K., and Rubin, K. Type I collagen synthesis in cultured

human fibroblasts: regulation by cell spreading, platelet-derived growth factor and interactions withcollagen fibers. Matrix Biol, 16: 409-425., 1998.

223. Wiig, H., Tveit, E., Hultborn, R., Reed, R. K., and Weiss, L. Interstitial fluid pressure in DMBA-induced rat mammary tumours. Scand J Clin Lab Invest, 42: 159-164., 1982.

224. Liotta, L. A. and Kohn, E. C. The microenvironment of the tumour-host interface. Nature, 411: 375-379., 2001.

225. Wernert, N. The multiple roles of tumour stroma. Virchows Arch, 430: 433-443., 1997.