Aspectos celulares y moleculares del cáncer

Clinical, Cellular, and Molecular Aspectsof Cancer Invasion

MARC MAREEL AND ANCY LEROY

Laboratory of Experimental Cancerology, Department of Radiotherapy and Nuclear Medicine,Ghent University Hospital, Ghent, Belgium

I. Introduction 338II. Cancer Pathogenesis 338

III. Invasion Promoter and Suppressor Genes 340A. The E-cadherin gene CDH1, an invasion/tumor suppressor 341B. The N-cadherin gene CDH2, an invasion promoter 343C. The E-catenin gene CTNNA1, a differentiation promoter 344D. The -catenin gene CTNNB1, a tumor/invasion-promoter gene 344E. Kinase and phosphatase genes, invasion promoters and invasion suppressors 346F. Invasion genes and metastasis genes, separate classes? 348G. Noncancer invasion-suppressor and invasion-promoter genes 349

IV. Cancer Cells, Host Cells, and Tumor Cells: All Invaders 349A. Cancer cells and host cells 349B. Myofibroblasts: stimulators of invasion 350C. Angiogenesis before invasion 350D. Tumor-infiltrated leukocytes: helpers of invasion 351E. Osteoclasts: targets for therapy 352F. Molecular cross-talk in noncancerous situations 352

V. Cellular Activities Associated With the Invasive Phenotype 353A. Cell-cell adhesion 354B. Cell-matrix interactions 357C. Migration 358D. Proteolysis 363

VI. Conclusions and Perspectives 365

Mareel, Marc, and Ancy Leroy. Clinical, Cellular, and Molecular Aspects of Cancer Invasion. Physiol Rev 83:337376, 2003; 10.1152/physrev.00024.2002.Invasion causes cancer malignancy. We review recent data aboutcellular and molecular mechanisms of invasion, focusing on cross-talk between the invaders and the host. Cancerdisturbs these cellular activities that maintain multicellular organisms, namely, growth, differentiation, apoptosis,and tissue integrity. Multiple alterations in the genome of cancer cells underlie tumor development. These geneticalterations occur in varying orders; many of them concomitantly influence invasion as well as the other cancer-related cellular activities. Examples discussed are genes encoding elements of the cadherin/catenin complex, thenonreceptor tyrosine kinase Src, the receptor tyrosine kinases c-Met and FGFR, the small GTPase Ras, and the dualphosphatase PTEN. In microorganisms, invasion genes belong to the class of virulence genes. There are numerousclinical and experimental observations showing that invasion results from the cross-talk between cancer cells andhost cells, comprising myofibroblasts, endothelial cells, and leukocytes, all of which are themselves invasive. In bonemetastases, host osteoclasts serve as targets for therapy. The molecular analysis of invasion-associated cellularactivities, namely, homotypic and heterotypic cell-cell adhesion, cell-matrix interactions and ectopic survival,migration, and proteolysis, reveal branching signal transduction pathways with extensive networks between indi-vidual pathways. Cellular responses to invasion-stimulatory molecules such as scatter factor, chemokines, leptin,trefoil factors, and bile acids or inhibitory factors such as platelet activating factor and thrombin depend onactivation of trimeric G proteins, phosphoinositide 3-kinase, and the Rac and Rho family of small GTPases. The roleof proteolysis in invasion is not limited to breakdown of extracellular matrix but also causes cleavage of proinvasivefragments from cell surface glycoproteins.

Physiol Rev83: 337376, 2003; 10.1152/physrev.00024.2002.

www.prv.org 3370031-9333/03 $15.00 Copyright 2003 the American Physiological Society

on July 8, 2014

Dow

nloaded from

I. INTRODUCTION

Cancer is malignant because cancer cells invade intoneighboring tissues and survive in this ectopic site. Theterm invasion indicates penetration into neighboring ter-ritories and their occupation. Cancer cells invade beyondthe constraints of the normal tissue from which theyoriginate; this invasion permits them to enter into thecirculation from where they can reach distant organs andeventually form secondary tumors, called metastases. In-vasion and metastasis are not unique for cancer as theyalso occur during embryonic development, in healthyadult organisms, and in many noncancerous diseases.Ectodermal cells invade through the primitive streak andoccupy the subectodermal space where they form themesoderm (118, 248, 252, 399, 410). Neural crest cellsemerge from the dorsal aspect of the neural tube, migrateto different sites in the body, and survive there to growand differentiate under the influence of specific localfactors (230, 400). Leukocytes leave their tissue of originin the bone marrow, enter into the circulation, and homeat specific sites (376). Microorganisms enter their host byinvasion through the lining epithelia of the skin or thegastrointestinal or respiratory tracts; they eventuallyreach the circulation and produce secondary lesions(346). We have discussed previously similarities in themolecular mechanisms of invasion by various organismstaking as examples four families of molecules, namely,cadherins, integrins, hydrolases, and chemokines (225).More recently, evidence was published in favor of a rolefor bacteria like Helicobacter pylori in the progression ofgastric cells toward a more invasive phenotype (170).

Noninvasive tumors are benign, because they arecured easily by simple removal. Invasive tumors, calledcancer, invariably kill their host if untreated and, evenwith optimal treatment, such tumors are a frequent causeof death (105). In the case of primary brain tumors, suchas astrocytomas, death is due almost uniquely to localinvasion, since, for yet unknown reasons, these tumorsrarely form metastases. Cancers of the head and neck,originating in the mucosa of the upper respiratory andalimentary tracts, kill mainly through local invasion andmetastasis to locoregional lymph nodes. Death by colo-rectal cancer is due to locoregional spread in one half andto distant metastasis in the other half of patients. In thecase of breast cancers and melanomas, death is usuallythe consequence of distant metastasis. It is quite obviousfrom the course of these diseases that invasion and me-tastasis are the hallmarks of cancer malignancy. Conse-quently, invasion and metastasis are major prognosticmarkers. The 5-year survival rate for bladder cancer thatis limited to the epithelium is 6080%, compared with3060% when invaded into the deeper muscle and 1040% when invaded through the bladder wall into the fat,provided radical treatment is performed. Melanoma 10-

year survival rates are 80% when the tumor invades intothe dermis but is 1.5 mm thick and 40% when itinvades into the subcutaneous fat. For primary brain tu-mors, the prognosis depends on the loss of differentiationrather than on the degree of invasion. Patients with well-differentiated astrocytomas, grade I and II, survive for 10years or more, whereas poorly differentiated astrocyto-mas, grades III and IV, have a median survival of 1 year.

In infectious diseases, though they are treated muchmore succesfully than malignant tumors, spread throughinvasion and metastasis may also herald a bad prognosis.In listeriosis, a disease caused by the bacterium Listeriamonocytogenes, spread from the site of entry in the intes-tine, to the meninges, or to the fetus in pregnant womencauses a potentially fatal disease (354). Resident nonin-vasive Streptococcus viridans is harmless; invasionthrough wounded oral mucosa and metastasis to dam-aged heart valves is a cause of death in 21% of the patients(211). Similarly, the cystic noninvasive form of Entam-oeba histolytica is harmless; invasion of E. histolyticatrophozoites into the enteric mucosa causes amoebic en-teritis with eventual spread to liver and brain (321).

Earlier experimental investigation of invasion andmetastasis focused on the development of appropriatemodels to score invasion and invasion-related activitiesby morphological techniques (1, 102, 109, 117, 164, 355,416, 448). Today, the study of invasion benefits from theenormous advances in genomics and proteomics, provid-ing thousands of genes and proteins, all well character-ized structurally and functionally. We review here themore recent data about cellular and molecular mecha-nisms of invasion, taking selected examples with empha-sis on homotypic and heterotypic cross-talk between theinvaders and the host. Noncancer invasion by normalcells or by prokaryotic and eukaryotic cells will be dis-cussed for comparison with cancer invasion. For a reviewof the older literature, the reader is referred to Reference249. A selected literature search for the present reviewwas closed in November 2001.

II. CANCER PATHOGENESIS

Cancer disturbs the cellular activities that are crucialfor the development and the maintenance of multicellularorganisms, namely, growth, differentiation, programmedcell death, and tissue integrity. Clinically, cancer mani-fests itself through a tumor because of excessive growth,through pain and bleeding because of invasion into nervesand vessels, and through functional disturbances becauseof pressure on and replacement of normal tissues. Thesesymptoms are not cancer specific, and the diagnosis ismade by histological examination of a sample from thetumor. This diagnosis includes the origin and type ofcancer, its extent of growth and invasion, and its grade of

338 MARC MAREEL AND ANCY LEROY

Physiol Rev VOL 83 APRIL 2003 www.prv.org

on July 8, 2014

Dow

nloaded from

differentiation. Attention is paid also to the host cellreaction evidenced by the stroma, blood vessels, andleukocytes. Because cancers are known to metastasize,the physician will search for secondary tumors in thelymph nodes and in distant organs. Growth, at least to theminimum volume detectable by the actual diagnostictechniques, is a prerequisite to find secondary tumors.Qualitative and quantitative criteria are used to stage andgrade cancers for therapeutic and prognostic purposes.Staging of tumors is done following the volume of theprimary tumor and its depth of invasion (T stage), thenumber and the volume of occupied lymph nodes as wellas invasion through their capsula (N stage), and the pres-ence of distant metastases (M stage). This TNM system,propagated by the International Union Against Cancer, iswidely used in Europe (372). For example, a T4, N1biii,M1 breast cancer has invaded into the skin, occupiedaxillary lymph nodes with invasion through their capsula,and has metastasized to distant organs such as bone, liver,brain, or lungs. A T2, N0, M0 breast cancer has a diameternot exceeding 5 cm and no metastasis are detected.Kaplan-Meyer survival curves show that patients withsuch T2 cancers, provided accurate treatment is given,have 70% chances of being alive 5 years after diagnosis.Attempts are made to refine staging by the identificationat diagnosis of these 30% of cancers that are not con-trolled by this treatment. The actual efforts include thesearch for micrometastases that are not detected in theroutine TNM system. Immunohistochemistry, with anti-bodies against epithelial marker proteins, of lymph nodesand bone marrow in breast cancers that were scored N0,M0 by the standard criteria showed immunopositive cellsin the lymph nodes in 6.4%, in the bone marrow in 26.0%,and in both in 4.8% of the patients (144). It is tempting tospeculate that the presence of such micrometastasesmight predict the above-mentioned 30% of lethal cases.One decade of careful follow up is, however, needed toknow the answer to the crucial question whether or notsuch nests of cancer cells will survive, grow, reach clini-cally relevant volumes, spread to other organs, and even-tually kill their host. Grading is based on the loss ofdifferentiation, sometimes combined with mitotic activ-ity. Individual cancers are currently portrayed by DNA,RNA, and protein microarray systems, covering as manycharacteristics of maligancy as possible. Whether or notthis method will enter into clinical routine for staging andgrading is an open question (194).

Noninvasive precursor lesions, i.e., histological ab-normalities in which cancer is more likely to occur than inthe normal tissue counterpart, are found in the vicinity ofinvasive cancers (71, 135). There is compelling evidenceto accept that many types of cancers have benign precur-sor lesions, recognized by accumulation of cells, such asin hyperplasia and adenoma, or by loss of differentiationand nuclear abnormalities, such as in atypia and carci-

noma in situ. There has been a long debate about whethera common initiated progenitor cell population would giverise to both noninvasive and invasive lesions (field theory)or a noninvasive precursor lesion would transit towardinvasive lesions (progression theory). In favor of the pro-gression theory is the concept that somatic mutationsfavoring continuous proliferation or low apoptosis led toclonal expansion and to continuous selection of progres-sively more malignant cell populations (334). Moreover,at least part of the genetic abnormalities of invasive can-cers are also found in apparently normal and in preinva-sive lesions, as exemplified in the breast (101), the pros-tate (100), the esophagus (199), and the bronchus (57).The scenario inferred from these clinical observations isconfirmed in models of experimental carcinogenesis inthe rat colon and in the mouse skin (384, 454).

The above-mentioned clinical and experimental ob-servations indicate that cancer is a disease of growth,causing accumulation of cells, of differentiation, causingloss of structure and function, and of tissue organization,leading to invasion and survival in an ectopic environ-ment (Fig. 1). A multistep process of invasion leads tometastasis: invasion from the tissue, in which the cancerhas originated, into the surrounding tissues through bar-riers such as the epithelial basement membrane; entryinto blood or lymph vessels; transport through the circu-lation; arrest and exit from the circulation at the putativesite of metastasis; and invasion into the tissues of theoccupied organ. Although cancer is generally thought toevolve from bad to worse, large variations in the rate ofprogression have been published. In a number of studiescollected from the literature about patients with cervicallesions that remained untreated against advice but ac-cepted follow up, progression from carcinoma in situtoward invasive carcinoma varied between 3 and 70%(71). For earlier stages of cancer development, the prob-ability of progression is lower than for more advancedstages of the disease, as suggested in Barretts esophagus(162). Such differences in progression between earlierand later stages of cancer development suggest a point ofno return where precursor lesions transit into lesions thatin most cases progress toward invasive cancer. This iscertainly the case for most cancers that have reached thestage of invasion and metastasis. There are, nevertheless,observations indicating that reversion to a more normalstage of at least part of the cancer cell population ispossible. Regression of metastasis from hypernephromaupon removal of the primary cancer without adjuvanttherapy has been described but remains a rare event withan unknown biological mechanism (105). Metastasessometimes show a higher degree of differentiation andgrow with less or no invasion compared with the primarycancer. This suggests that the progression model shouldnot assume that invasion and metastasis-associated phe-notypes are fixed by genetic alteration. Colorectal carci-

CANCER INVASION 339


on July 8, 2014

Dow

nloaded from

noma metastases may resemble the organized epithelialand tubular structure of a well-differentiated primary can-cer, whereas the invasive front of the actual primarycancer displays loss of the epithelioid morphotype andappearance of fibroblastoid, presumably invasive andmetastatic, cancer cells. An experimental demonstrationis provided by cocultures of human colon cancer cellswith enteric lymphoid cells, in which the cancer cellstransit to M cells, a differentiated type of absorptive en-terocytes covering Peyers patches (208).

There are indications that abnormal growth and in-vasion are not necessarily associated during cancer de-velopment. Metastasis without primary tumor, called CUPfor cancer with unknown primary, is not a rare event(510% of all cancer patients with metastases). In suchcases, the primary tumor cannot be found at the time ofclinically evident, hence growing, metastases. Primarytumors may appear later or not at all. Conversely, metas-tases may grow to reach clinically relevant volumes manyyears after removal of the primary cancer as exemplifiedby ocular melanoma. Both clinical observations indicatethat invasion and growth at the primary or secondary sitecan be regulated independently, a conclusion that is con-firmed by the experimental finding that pharmacologicalagents can arrest growth whilst permitting invasion, andvice versa (385). This growth-separate-from-invasion con-cept deals probably with the exception; in the majority ofinvasive cancers, a complex and probably coordinatedprogram of invasion, growth, survival, and loss of differ-entiation is at the basis of the clinical manifestations ofthe disease. It is, indeed, logical to accept that prolifera-tion is needed to provide a cohort of invaders and thatinhibition of apoptosis keeps them alive in an ectopicmatrix environment. Considering noncancer invasion,leukocytes do proliferate in the bone marrow and, there-

after, invade and metastasize as nondividing cells. Para-sitic trypanosomes and leishmania pass through an oblig-atory nondividing stage when they invade from one hostinto another or from one tissue into another (279).

III. INVASION PROMOTER ANDSUPPRESSOR GENES

A series of alterations in the genome of the cell popu-lation of origin forms the basis of tumor development (40,130). The genes of interest are classified as oncogenes ortumor-promoter genes, one allele of which is activated lead-ing to gain-of-function events, and tumor-suppressor genesor antioncogenes, both alleles of which are inactivated lead-ing to loss-of-function events. Genomic instability, due eitherto impairement in DNA repair (microsatellite instability) orto dominant negative mutations in mitotic check-point genes(chromosomal instability), leads to activation of oncogenesand inactivation of tumor suppressor genes. The products ofthese genes belong to various classes of protein families,such as cytokines, cell surface receptors, signal transducers,and transcription factors. The list of oncogenes encodingcell surface receptors of the protein-tyrosine kinase familyalone counts more than 40 members (42). Mechanisms ofactivation of oncogenes implicate mutation, gene amplifica-tion, and promoter activation. Mechanisms of tumor-sup-pressor inactivation are exemplified by loss of heterozygos-ity (LOH) plus silencing of the second allele genetically,through mutation, or epigenetically, through methylation. Infamilial cancers, one mutation is carried with the germline.Well-documented examples include RB in retinoblastoma,BRCA1 in breast cancer, and adenomatous polyposis coli(APC) in colon cancer of the familial adenomatous polypo-sis (FAP) type. Cancer-related genetic alteration are multiple

FIG. 1. Schematic representation of genetic,epigenetic, and phenotypic aspects of cancer de-velopment. [Adapted from Mareel and Bracke(247).]



on July 8, 2014

Dow

nloaded from

and occur in varying orders so that it is difficult to ascribedefined genetic alterations to distinct stages of tumor devel-opment (18, 77, 456). The sequence of genomic alterationsduring tumor development might be of particular interest forthe understanding of their role in the acquisition of invasion.Are the genes implicated in invasion different from thoseimplicated in growth disturbance, loss of differentiation, andof sensitivity to death signals? The recognition of some stagespecificity of genetic alterations has led us previously tobelieve that oncogenes and tumor-suppressor genes wereimplicated in growth disturbance and that they differed fromgenes promoting or suppressing differentiation, invasion, orsurvival (250). The growing list of cancer genes, however,comprises several examples of oncogenes and tumor-sup-pressor genes that are, either in the same or in differenttypes of tumors, implicated in earlier stages of growth dis-turbance as well as in the later stages when invasiveness isacquired (27). Furthermore, the categorization of tumor phe-notypes in growth, differentiation, survival, and invasionunderestimates the mutual relationships between these phe-notypes (125). To illustrate, growth is the basis for clonalexpansion of somatic cells, differentiation and proliferationare inversely related, and survival signals are implicated ininvasion because normal cells die inside foreign trophicmicroecosystems (358). The known function of the geneproduct, sometimes, makes the stage of appearance of thegenetic alteration unexpected. Loss of p53, the guardian ofthe genome, is frequently found at the transition betweenthe noninvasive, premalignant and the invasive, potentiallymalignant stage, and this is later than expected from themore general role of the p53 phosphoprotein functioning inthe check-point control that arrests cells with damagedDNA. In line with this observation, p53 null mice are lesssusceptible to induction of papillomas, but once the papil-lomas arise, they transit rapidly to carcinomas. The effect ofgene activation may depend on the stage of development atwhich it occurs. For example, transforming growth factor-(TGF-) acts as a tumor suppressor in early stages of tumordevelopment, whereas it causes invasion and metastasisupon inducible transgenic expression in papillomas (439).The following discussion about invasion-suppressor and in-vasion-promoter genes and their alterations during tumordevelopment chooses examples on the basis of the interestof the authors laboratory. The more tumor-suppressor or-promoter genes are examined, the better it is realized thatmany of them affect invasion as well as growth and differ-entiation.

A. The E-cadherin Gene CDH1,an Invasion/Tumor Suppressor

Epithelial (E)-cadherin is a transmembrane glycopro-tein of the type I cadherin superfamily (299); its cytoplas-mic part is linked to the actin cytoskeleton via the

catenins, -catenin, -catenin, and plakoglobin (-cate-nin). The gene encoding E-cadherin (CDH1, on chromo-some 16q22.1) was one of the first to be considered as aninvasion-suppressor gene (30, 138, 432). The experimentalstrategy consisted of the isolation from heterogeneouscell lines of clones with an epithelioid (e-type, resemblingepithelial cells) morphotype and a fibroblastic (f-type,resembling fibroblasts) morphotype. The e-type cellswere E-cadherin positive, failed to invade into organotypi-cally cultured embryonic chick heart, and formed a dif-ferentiated epithelial layer around the heart tissue. Thef-type cells were E-cadherin negative, did invade, andshowed no epithelial differentiation. Similarly, a positivecorrelation was found between the invasion into collagentype I of human cancer cell lines and the lack of E-cadherin. The invasive phenotype as well as the morpho-type of these cells could be manipulated in both direc-tions, from e-type noninvasive to f-type invasive, and viceversa, by transfection with sense or antisense E-cadherincDNA. The e- to f-type conversion is reminiscent of theepithelial to mesenchymal transition (coined EMT) ob-served during gastrulation. Interestingly, loss of E-cad-herin in immortalized cell lines of noncancerous origindid induce the invasive phenotype, only when the cellswere transfected with an oncogene (Fig. 2). Conclusionsfrom these experimental findings were confirmed by im-munohistochemical changes in E-cadherin expressionand localization in most human cancers (56, 92, 98, 275,276, 305, 353, 366, 419, 421). The positive correlationbetween cancer aggressiveness as evidenced by poor sur-vival and disturbance of E-cadherin provides clinical sup-port for E-cadherin as an invasion suppressor (304). Suchclinical evidence is, however, at best circumstantial sincedeficient expression of E-cadherin may also be due toposttranscriptional and posttranslational events. Thecausal relationship between E-cadherin expression andinvasion in vivo was convincingly demonstrated in trans-genic mice (319). Such mice, expressing the tumorigenicsimian virus 40 (SV40) T antigen under the insulin pro-moter (Rip1Tag), developed pancreatic -cell adenomasin 74% and invasive adenocarcinomas in 26% of the mice.Overexpression of E-cadherin under the same promoter(Rip1E-cad) did not cause tumors. Rip1Tag Rip1E-cadcrosses had a lower (8%) ratio of adenocarcinomas, show-ing that overexpression of E-cadherin counteracted theacquisition of the invasive phenotype. In contrast, crossesof Rip1Tag mice with mice expressing dominant negativeE-cadherin that lacks the extracellular domain (Rip1dnE-cad) had invasive and metastatic carcinomas in 50% of thecases.

Mutations in CDH1 are the exception rather than therule as they occur only in diffuse type gastric cancer,lobular breast cancer, and endometrial cancer (2527, 35,36, 196). In invasive lobular breast cancers, a subtype inwhich cancer cells invade as Indian files, total loss of

CANCER INVASION 341


on July 8, 2014

Dow

nloaded from

E-cadherin expression is due to E-cadherin gene muta-tions combined with loss of the wild-type allele. Theabove-mentioned observations are compatible with inac-tivation of CDH1 either at the transition between thenoninvasive to the invasive stage or earlier. In lobularbreast cancer early inactivation of CDH1 was demon-strated, putting forward CDH1 also as a tumor-suppressorgene (435). The same truncating mutations associatedwith loss of heterozygosity were found in sporadic lobularcarcinoma in situ as in the associated invasive compo-nents but not in atypical hyperplasia. In diffuse type gas-tric cancers, the amplification product of E-cadherincDNA was shorter than the expected 630 bp due to skip-ping of exon 9 or 8. In favor of the tumor-suppressorfunction of the E-cadherin gene is the finding of inactivat-ing germ line mutations in families with a higher inci-dence of diffuse gastric cancer. Interestingly, the motherof one of the gastric cancer patients suffered from meta-chronous lobular breast cancer and diffuse gastric cancerand had the same CDH1 germline mutation as her child(206).

The low frequency of E-cadherin mutations is instriking contrast to the almost ubiquitous disturbance ofE-cadherin in invasive and even in preinvasive cancers,suggesting other mechanisms of transcriptional or post-transcriptional downregulation. Such forms of downregu-lation may be reversible. In an experimental tumor modelwith an immortalized normal kidney-derived epithelialcell line Madin-Darby canine kidney (MDCK) transformedby a mutated RAS oncogene and coined MDCK-ras, E-cadherin-positive variants had an e-morphotype and werenoninvasive in vitro, but produced invasive and meta-static cancers after injection into nude mice (246). Immu-nohistochemistry of the nude mouse tumors revealed lossof E-cadherin, but ex vivo culture of the MDCK-ras tu-mors resulted in rapid reexpression of E-cadherin, acqui-

sition of an e-morphotype, and loss of invasiveness. Thehost mouse context responsible for the changes in E-cadherin expression has, not yet, been identified. Revers-ible downmodulation of E-cadherin is suggested also byits reexpression in metastases from breast cancers (63).Similar observations with colorectal cancers led to theconclusion that, to grow at the metastatic site, dissemi-nated f-type cancer cells must regain at least some of theirepithelial functions (53).

Methylation of DNA is a common type of transcrip-tional modification in mammals. It normally occurs duringgenomic imprinting and X chromosome inactivation.Highly methylated DNA is found in genetically silent re-gions of chromosomes, and hypermethylation of CpG is-lands in the promoter region of a gene leads to transcrip-tional silencing. This mechanism of downregulation wasobserved in several tumor-suppressor genes such as APC,von Hippel-Lindau (VHL), RB, and also CDH1 (124).Germline mutations in CDH1 without loss of heterozygos-ity at the CDH1 locus are suggestive for hypermethyl-ation. In hereditary diffuse gastric cancer, with a mutationin one CDH1 allele, hypermethylation constitutes the sec-ond hit eliminating the expression of E-cadherin (155). Inthis type of cancer and also in esophageal cancer, hyper-methylation may occur as early as the intramucosal, i.e.,noninvasive stage, of the disease, underscoring the tumor-suppressor function of E-cadherin (116, 391). In a gastriccancer cell line, the expression of E-cadherin could berestored by treatment with the demethylating agent 5-aza-cytidine. Some authors have cautioned interpreting hy-permethylation in terms of tumor development, as theyconsider that the causal relationship between both phe-nomena is not firmly established (129).

The promoter of CDH1 contains positive regulatoryelements, a CCAAT-box and GC-boxes, as well as twoE-boxes (29, 150). Proteins acting directly or indirectly on

FIG. 2. E-cadherin promoter withpositive and negative transcriptional reg-ulators. For detailed description, see sec-tion IIIA. [Adapted from Van Aken et al.(419).]



on July 8, 2014

Dow

nloaded from

the E-cadherin promoter are presented in Figure 2. Notethat some of these proteins are encoded by genes thatwere classified as tumor-suppressor genes or as protoon-cogenes. In MDCK cells, coined MDCK(LT), the SV40large-T antigen, encoded by a viral oncogene, inactivatesRB and causes a transition from the e- to the f-morpho-type that is associated with loss of epithelial markers,including E-cadherin, and with loss of expression of theoncogene MYC (258). The latter encodes two distinct Mycproteins, acting as transcription factor (22). Transfectionsof dog kidney MDCK and human skin HaCat cells showthat RB and Myc specifically activate transcription of theE-cadherin promoter, a phenomenon that is mediated bythe transcription factor AP-2 (23). Transactivation of theE-cadherin promoter is strongly dependent on the expres-sion ratio between the two Myc proteins and is cell typespecific. The finding that inactivation of RB by humanpapilloma virus HPV16E7 in primary human mammaryepithelial tissue explanted in reconstituted extracellularmatrix does not interfere with the correct expression ofE-cadherin confirms this cell type specificity (374). Insuch cultures, viral inactivation of RB causes loss of thedifferentiation markers lactoferrin and cytokeratin-19.These experiments illustrate the complexity of oncogenicpathways: a viral tumor-promoter binds to and inactivatesa cellular tumor-suppressor that indirectly counteractsinvasion in one type of cells and maintains differentiationin another type. The Wilms tumor 1 (WT1) may transac-tivate CDH1 directly through binding to the proximalGC-rich sequence in the promoter as evidenced by trans-fection of 3T3 fibroblasts (174). Snail, a member of amulti-zinc finger protein family of transcription factors, isa strong repressor of CDH1, interacting specifically withE-boxes in the E-cadherin promoter and, so, repressingtranscription of CDH1; in this way, it causes an e- tof-morphotype transition and invasion (21, 68). Snail isexpressed in fibroblasts, in some E-cadherin-deficient celllines, and in invasive regions of experimental carcinomas.Smad interacting protein 1 (SIP1) belongs to the samezinc finger protein family as Snail and displays specificDNA binding activity (81). It interacts with several mem-bers of the Smad protein family. Conditional expressionof SIP1 in MDCK-Tetoff cells, an MDCK-derivative stablyexpressing the Tetoff transactivator, abrogates the ex-pression of E-cadherin and of cell-cell adhesion as well asunidirectional migration that are both sensitive to inhibi-tion by E-cadherin-neutralizing antibodies; it simulta-neously induces invasion into collagen gels. Members ofthe Smad protein family, which normally act in the TGF-signaling pathway cooperatively with other transcriptionfactors (104, 450), are implicated in invasion in, yet, an-other way. SMAD2 genes with a mutation of Ser at posi-tion 465 are found in colon cancer and in lung cancer.When such genes, encoding an unphosphorylable form ofSmad2, are transfected into MDCK cells or into human

colon cancer HCT-8 cells, coined HCT-8/E11 for clonalselection of an epithelioid morphotype, they induce theinvasive phenotype, and invasion is enhanced by additionof TGF- to the culture medium through an, as yet, un-known mechanism (328). In cells from the normal murinemammary gland (NMuMG) family (273) and in pancreaticcancer cells carrying an activating RAS mutation (121),TGF- caused an e- to f-morphotype transition and inva-sion, with downregulation of E-cadherin and of otherjunctional proteins.

B. The N-cadherin Gene CDH2,an Invasion Promoter

Gain of N-cadherin in cancer cells accompanies lossof E-cadherin, acquisition of an f-morphotype, increasedmotility, and invasion both in vitro and in vivo as summa-rized in Reference 419. In some of these cancer cells, E-and N-cadherin are coexpressed. In such cells, the inva-sion-promoter potency of N-cadherin seems to dominatethe invasion-suppressor potency of E-cadherin. Indeed,N-cadherin promotes invasion and motility of humanbreast cancer cells in a way that is not overcome byforced expression of E-cadherin (290, 291). A weaklymetastatic E-cadherin expressing breast cancer cell-lineof the MCF-7 family, yielded, upon successful transfectionwith N-cadherin, cells that coexpressed E- and N-cadherinand that were highly metastatic (167). Conversely, E-cadherin transfection in N-cadherin expressing breastcancer cells did not revert their invasive phenotype. Theshifts from E-cadherin to N-cadherin raise the questionwhether the expression of both genes is coregulated.Transfection experiments yielded conflicting results. De-creased N-cadherin expression upon transfection withcDNA encoding L-CAM, the chicken homolog of E-cad-herin, was ascribed to instability of the N-cadherin pro-tein and not to reduced transcription (234). In squamouscarcinoma cells, transfection of N-cadherin cDNA causeda decrease in E-cadherin expression; conversely, whenN-cadherin expression was decreased by antisense trans-fection, E-cadherin expression increased (182). That re-pressors of E-cadherin may transactivate the gene encod-ing N-cadherin is demonstrated for the zinc finger proteinSnail (21, 68). The retinal pigment epithelium (RPE) of thehuman eye may provide an interesting experimentalmodel for the further analysis of the E- to N-cadherin shift(64; E. Van Aken, personal communication). In the eye,the RPE expresses mainly E-cadherin; shortly after ex-plantation in vitro, RPE cultures show a majority of E-cadherin-negative, N-cadherin-positive cells with an f-morphotype. When such RPE cells are seeded on collagengel, they extensively invade, and invasion can be blockedby addition of the N-cadherin neutralizing antibody GC-4.Clinical data do not substantiate unanimously the inva-

CANCER INVASION 343


on July 8, 2014

Dow

nloaded from

sion promoter role of N-cadherin. Immunohistochemicalanalysis of 75 bladder cancers led to the conclusion thatfocal expression of N-cadherin in urothelial cancers is afrequent phenomenon, but its significance for invasion isunclear (335).

C. The E-catenin Gene CTNNA1,a Differentiation Promoter

-Catenins are considered to be essential elements ofthe E-cadherin invasion suppressor complex (173, 415).The CTNNA1 gene encoding E-catenin is localized onchromosome 5q3.1 (300). In cultures of the human coloncancer clone HCT-8/E11, round cells (r-morphotype) canbe observed either on top of the epithelioid (e-type) celllayer or floating in the culture medium (429, 431). Exam-ination of the E-cadherin/catenin complex in harvested orcloned r-type cells showed that loss of E-catenin causedthe transition from an e- to r-morphotype in an irrevers-ible manner. Repeated screening of a series of coloncancer cell lines stored in our and in other laboratories aswell as purchased from commercial stock, showed a sim-ilar morphotypic instability with spontaneous emergenceof similar r-type variants in routine culture as observedwith HCT-8/E11. DNA fingerprinting of cell lines coinedHRT-18, DLD-1, or HCT-15 all had the same microsatelliteinstability DNA profile as HCT-8, strongly indicating thatthey all originated from the same patient (430). Suchconfusion about the identity of cell lines kept in variousdeposits was recently estimated to be 36% (262). In thecase of the HCT-8 family, the confusion was useful be-cause it pointed toward a genetic background for the e- tor-type transition as it occurred in all the cell lines that hadthe same DNA profile. Cells of the HCT-8 family carry aheterozygous mutation in a CTNNA1 gene and mutationor loss of the remaining wild-type allele causes loss of-catenin and the above-mentioned e- to r-morphotypetransition. CTNNA1 is, therefore, considered as a tumor-suppressor gene in accordance with Knudsons criteria(213). The conclusion about the cellular activity sup-pressed by E-catenin depended on the assay used to testthe cells. In the chick heart invasion assay (Fig. 3), E-catenin-positive e-type HCT-8/E11 cells are not or poorlyinvasive with a few undifferentiated cells inside the pe-ripheral rim of the heart tissue, in contrast to the E-catenin-negative r-type cells that massively occupy theheart tissue without any sign of differentiation. Theseexperiments led to the conclusion that CTNNA1 is aninvasion-suppressor gene (429). However, differentiationconstituted another striking difference between the twocell types. When assayed on top of collagen type I gels,both types of cells fail to invade unless cancer-associatedmyofibroblasts are admixed to the collagen (106). In thepresence of myofibroblasts, both e- and r-type cells do

invade, the former as epithelioid strands and the latter asloose files of undifferentiated cells. It is, therefore, justi-fied to consider CTNNA1 as a differentiation gene, influ-encing invasion only in a quantitative way and, possibly,as a consequence of changes in differentiation. Wheninjected orthotopically into the wall of the cecum of nudemice, genotypic differences in CTNNA1 between e- andr-type cells were conserved, but phenotypic differencescould not be seen any more; both variants producedmoderately differentiated invasive adenocarcinomas thatwere undistinguishable from one another (426). Reex-pression of -catenin in an E-cadherin-positive prostatecancer cell line PC-3 suppressed tumorigenicity in nudemice (127), suggesting a role for -catenin in ectopicsurvival. Recent data suggest that E-catenin not onlyfunctions through maintainance of cell-cell adhesion butalso through interference with -catenin/Tcf/DNA com-plex formation and -catenin signaling in the nucleus(147). The above-mentioned data again illustrate the im-plication of a single gene in multiple tumor progression-associated phenotypes.

D. The -catenin Gene CTNNB1,a Tumor/Invasion-Promoter Gene

-Catenin, like the other catenins, was described firstas an essential element of the E-cadherin/catenin com-plex (311). In normal cells, -catenin is associated notonly with cadherins but also with the APC multiproteincomplex (157). Here, it is phosphorylated by the serine/threonine kinase glycogen synthase kinase (GSK)-3 anddirected to the ubiquitine proteasome pathway for degra-dation (see Fig. 8). The APC complex belongs to the Wntsignaling pathway, in the context of which -catenin mayact as a tumor promoter (32, 341, 387, 446). Mutations inthe serine/threonine phosphorylation sites of -cateninmake it resistant to degradation. After saturation of theE-cadherin complex, the superfluous -catenin stays inthe nucleus in association with lymphocyte enhancer fac-tor (LEF)/T-cell factor (TCF) influencing transcription.Because TCF proteins possess no intrinsic ability to mod-ulate transciption, coactivators such as the acetyltrans-ferases p300 (169), Smad3 (220), and Pontin52 (24) arecrucial. In contrast, Reptin52 (repressing pontin52) actsas a repressor. An overview of -catenin mutations inhuman tumors is listed in Reference 298. This list containsdesmoid tumors, locally invasive outgrowths of mesen-chymal cells that do not produce metastases (398).-Catenin mutations are limited to certain types of can-cer; they are not found in squamous cell carcinomas ofhead and neck and esophagus, gastric carcinomas, orlobular and ductal breast carcinomas (67, 99, 153). In aseries of 58 colorectal cancers without APC mutation,there were no CTNNB1 point mutations, but 7 tumors



on July 8, 2014

Dow

nloaded from

showed deletions of 234760 bp, each of which includedall or part of exon 3 (184). Retention of -catenin in thenucleus may result also from mutations in the -cateninbinding site of APC, which is necessary for the formationof the complex in which GSK-3-mediated phosphoryla-tion occurs (283). Mutation in the armadillo repeats is oneof the oncogenic changes in APC, a tumor suppressor thatis implicated in sporadic and familial colon cancer (326),as evidenced also in transgenic mice (133). The participa-tion of APC at tumor development is probably not limitedto growth, as the APC proteome also regulates morpho-genesis (19). Target genes of the -catenin/TCF transac-tivator complex comprise the genes encoding matrix met-alloproteinase-7 (MMP-7; matrilysin) (90), myc (168),cyclin D1 (236), multidrug resistance protein 1 (MDR1)(453), two components of the AP-1 transcription complexjun and fra-1 (245), and the putative transcriptional regu-lator AF17 (237). In melanoma cells, the acquisition of theTGF--dependent fibroblastic morphotype is accompa-

nied by localization of -CTN in the nucleus and an in-creased expression of MMP-9, next to an increase ofintegrin-linked kinase (ILK), 1- and 3-integrins, and adecrease of E-cadherin (186). We would like to empha-size, here, that an element of an invasion-suppressor com-plex transactivates genes that are implicated not only inthe modulation of growth, differentiation and response totherapy, but also in the stimulation of invasion and inva-sion-associated cellular activities. In cancers where mu-tations in -catenin, APC or conductin (equal to axin) (28)are uncommon, -catenin may be upregulated by theprolyl isomerase (Pin1) resulting in the transcription ofseveral -catenin target genes (344). Pin1 interferes with-CTN/APC interaction by changing the conformation ofthe phosphorylated Ser/Thr-Pro bonds. There is less evi-dence in favor of CTNNB1 as an invasion-suppressorgene. An in-frame deletion in the E-catenin binding siteof -catenin in a signet ring cell carcinoma cell line (HSC-39) causes disruption of the cadherin-dependent cell-cell

FIG. 3. In vitro assays for invasion of cancer cells and of bacteria. For further details about these and other assaysfor cancer cell invasion, see Ref. 61; for the bacterial invasion assay, see Ref. 349.

CANCER INVASION 345


on July 8, 2014

Dow

nloaded from

adhesion (196, 201, 310). In normal mesenchymal cellsand in uveal melanoma, cytoplasmic nonphosphorylated-catenin might well participate at invasion without trans-locating into the nucleus (210).

Plakoglobin, also called -catenin, is a close homologof -catenin, sharing the 12 armadillo repeats; both maypromote tumor formation when overexpressed (17). Inadherens junctions, these two molecules bind indepen-dently to E-cadherin, and plakoglobin binds also to thedesmosomal cadherins desmoglein and desmocollin. Un-like for -catenin, neoplastic transformation by plakoglo-bin does not implicate stabilizing mutations (215). It does,however, require transcriptional activation of the onco-gene MYC via plakoglobin/TCF complexes. One possibledifference between the two armadillo proteins may be thetype of binding coactivator proteins as suggested by dif-ferences in their carboxy-terminal parts. In experimentalsystems, plakoglobin may also act as a tumor suppressorand as a tumor promoter (458). Overexpression of plako-globin in cells from a human tongue squamous cell carci-noma SCC9 causes uncontrolled growth and inhibition ofapoptosis. Here, plakoglobin exerts a growth regulatoryfunction by induction of the antiapoptotic protein BCL-2,independently of its role in mediating cell-cell adhesion(160). In the above-mentioned observations of metastaticbreast cancer by Bukholm et al. (63), most members ofthe E-cadherin/catenin complex, including -catenin,were reexpressed in the metastases, whereas plakoglobinwas lost. Mutation of APC, as frequently observed incolon cancer, results in elevated levels of both -cateninand plakoglobin (216, 283). Taken together, plakoglobinand -catenin are best considered as separate players intumor development with close links to each other as well asto other elements of the cadherin and the APC complexes.

E. Kinases and Phosphatase Genes, InvasionPromoters and Invasion Suppressors

Phosphorylation and dephosphorylation are key phe-nomena in intracellular signaling, and genes encodingkinases and phosphatases are on the list of oncogenes andtumor-suppressor genes. We discuss their putative rolesin invasion, taking the examples of the nonreceptor ty-rosine kinase SRC, the receptor tyrosine kinases c-METand FGFR, the small GTPase RAS and the dual phospha-tase PTEN. Doubtless, many more genes encoding othermembers of these protein families may act as invasionpromoters or invasion suppressors (see list in Ref. 42).

1. Src

SRC is the first oncogene detected (381) and a pro-totype showing many characteristics of the other onco-genes (259). The viral oncogene v-SRC has a cellularcounterpart c-SRC that is activated by mutation to be-

come an oncogene. Such activating mutations are foundin human cancer both in early and late stages of develop-ment (72, 180). In some colorectal cancers, SRC activa-tion, through truncating mutations in a critical carboxy-terminal tyrosine, probably has a role in malignantprogression (180), supporting its invasion-promoter func-tion. The Src protein is anchored to the plasma membranethrough myristoylation, receives various signals, signalsin its turn to many substrates directly and indirectly (seeFig. 11), and is implicated in numerous cellular functions,including proliferation, motility, as well as cell-cell andcell-substrate adhesion (402). Temperature-sensitive mu-tants of SRC were used to transform MDCK cells, theinvasion of which into embryonic chick heart or intocollagen gels could be switched on by changing the incu-bation temperature from 39.5 to 35C (31). In these MD-CKts.src cells, activation of Src leads to loss of E-cadherinfunctions, as evidenced by deficient cellular aggregationand gain of invasion. One of the molecular changes ofinterest in the E-cadherin/catenin complex is tyrosinephosphorylation of -catenin, weakening its binding withE-cadherin. In PC/AA/C1 cells, derived from a colon polypin a FAP patient, introduction of activated SRC was notsufficient to induce invasion, but made the cells sensitiveto stimulation of invasion by other factors (123). Indeed,the parental cells (PC/AA/C1) failed to invade into colla-gen in vitro, even after stimulation with scatter factor(SF)/hepatocyte growth factor (HGF), whereas their SRC-transfected derivatives (PCmsrc) did invade upon addi-tion of SF/HGF. The membrane-associated polyoma mid-dle T oncoprotein (py-MT), known to increase thetyrosine kinase activity of pp60c-src by preventing phos-phorylation of Y527, mimicked some of the transformingeffects of SRC but not its proinvasive activity (294). In ratbladder cells NBT-II, Src activity correlates with loss ofepithelial differentiation and metastasis (51). During gas-trulation movements in the Xenopus embryo, Src kinasesare pivotal as evidenced by in vivo injection of mRNAscoding for dominant negative forms of ubiquitous mem-bers of the Src family.

2. Receptor tyrosine kinases

The activation of receptor tyrosine kinases, such asc-Met and FGFR, contributes to invasion. The c-Met re-ceptor for SF/HGF consists of a 50-kDa extracellular-subunit that is disulfide-linked to a 145-kDa -subunithaving cytoplasmic tyrosine kinase domains and sites oftyrosine phosphorylation (48, 286). The effects of SF/HGFon motility, invasion, and proliferation are due to activa-tion of the c-Met receptor as shown by transfection ofhybrid cDNA encoding the ligand binding domain of nervegrowth factor and the transmembrane and tyrosine kinasedomains of c-Met (442). Mutation and overexpression ofc-Met are associated with tumor progression in various



on July 8, 2014

Dow

nloaded from

human cancers, including kidney (356), thyroid, pancreas,colorectal (110, 111) and gastric cancers (231). In hered-itary papillary renal cell carcinoma, missense mutationsin the MET protooncogene lead to constitutive activationof the c-Met protein (356). Activated multifunctionaldocking sites recognize SH2-containing adaptors like Grb2and Shc, attract effector proteins such as phosphoinosi-tide (PI) 3-kinase, Src, and phospholipase C (PLC)- andso stimulate diverse cellular functions (327). Phosphory-lation of Y1349 and Y1356 is essential for scattering; pointmutation in Y1349 abolishes the metastatic potentialwhilst enhancing the transforming activity. In NIH 3T3cells, transfected with a mutated c-MET, tyrosine phos-phorylation corresponds with transforming potential, fo-cus formation, and tumorigenicity (189). Mutant c-Metinduces motility in MDCK cells and metastatic potential inNIH 3T3 cells; transgenic mice develop metastatic mam-mary cancer (188). Interestingly, the expression pattern ofthe c-Met, and of other protooncogenic receptor tyrosinekinases, with their respective ligands during embryonicdevelopment suggests that they are involved in normalepithelial morphogenesis as well (39).

The FGFR family consists of four genes: FGFR-1 (flggene), FGFR-2 (bek gene), FGFR-3, and FGFR-4. In therat Dunning prostate cancer model and in human prostatecancer , progression is associated with alternative spicingin FGFR-2 gene: early stages express the III-b isoformdominantly binding keratinocyte growth factor (KGF) andlater stages the III-c isoform with high affinity for basicfibroblast growth factor (bFGF) (142). FGFR1 and FGFR2c/bek transfection into NBT-II cells leads to epithelial-mesenchymal transition in response to acidic fibroblastgrowth factor (aFGF) and bFGF (352). Similarly, trans-fection of FGFR-1 into less malignant prostate cells ac-celerated the progression toward a more maligant pheno-type (131, 263). It is interesting to note the existence of anFGF-2 nuclear isoform conveying metastatic propertiesupon NBT-II rat bladder carcinoma cells without secre-tion and without the conventional FGFR-mediated signal-ing pathway (306).

3. Ras

The RAS protooncogenes encode membrane-associ-ated guanine nucleotide binding proteins of 21 kDa. Con-stitutive activation by substitution of amino acid residuesat various positions is frequently found in human invasivecolorectal carcinomas and in other types of human cancer(45). Interestingly, activated Ras cooperates with TGF-to regulate invasion via the Raf-MAPK pathway (154, 185,303, 399). Invasion into chick heart as well as into colla-gen gels was conveyed by Py-MT or by mutated (Val-12)Ha-ras upon SLC-44 rat intestinal epithelial cells immor-talized by polyoma large T but not upon Caco-2 cells,derived from a human colonic adenocarcinoma cell line

(75). Interestingly, the SLC-44 cells and its derivatives hadweak or no expression of E-cadherin, whereas all Caco-2cells were clearly positive at the cell-cell borders. In theseexperiments the effect of constitutive expression of theoncogenes was not limited to invasion, since it also in-creased growth as measured after subcutaneous trans-plantation into nude mice. Moreover, Ha-RAS transfectedCaco-2 cells failed to perform enterocytic differentiation.Examples of the competition between the invasion-sup-pressor effect of E-cadherin and the invasion-promotereffect of activated Ras are shown in Figure 4. Clearly, inthese and in other model systems, the invasion-suppres-sor potency of E-cadherin did neutralize the invasion-promoter potency of the mutated Ha-Ras (432).

4. Pten

PTEN (phosphatase and tensin homolog deleted onchromosome 10), also termed MMAC1 (mutated in mul-tiple advanced cancers) or TEP1 (TGF- regulated andepithelial cell enriched phosphatase 1), was the first tu-mor suppressor gene encoding a protein with phospha-tase activity (69). The gene is located on chromosome10q23 and is a candidate invasion suppressor because ofits late inactivation during cancer development (233, 379,392). Its mutation frequency in human cancers is veryhigh and close to that of p53. Mutations were described inhigh-grade but not in low-grade gliomas, independent ofp53 mutations (114, 330), in bladder cancer (65), in ad-vanced prostate cancer, and cell lines derived therefrom(66, 444). PTEN acts on both proteins and lipids; as aprotein phosphatase it has a dual specificity acting ontyrosine and on serine/threonine. Its main targets as alipid phosphatase are phosphoinositides where PTEN de-phosphorylates the three position and in this way coun-teracts PI 3-kinase. Clustering of mutations in the lipidphosphatase domain, e.g., G129E, suggests that this do-main is critical for the tumor-suppressor activity (284).The involvement of the protein phosphatase domain andthe loss of expression observed in some cancers led to theconclusion that PTEN has a dual role, regulating growthand survival through its lipid phosphatase activity andadhesion and invasion through its protein tyrosine phos-phatase activity (392, 444). More recent experiments withRAS or SRC transformed MDCK (MDCK-ras andMDCKts.src) cells favor the opinion that the lipid phos-phatase activity of PTEN is implicated in stabilization ofjunctional complexes and restriction of invasion (218).Indeed, successful transfection of these MDCK transfor-mants with wild-type PTEN, but not with mutants defi-cient in lipid phosphatase activities, induces cellular ag-gregation and abolishes invasion. The implication of theE-cadherin/catenin complex in the expression of the non-invasive phenotype was demonstrated by the proinvasiveaction of antibodies that functionally neutralized E-cad-

CANCER INVASION 347


on July 8, 2014

Dow

nloaded from

herin. The same tranfections counteracted invasion alsoin PTEN-defective cell lines derived from neuroblastoma,melanoma, and prostate carcinoma. Like for the genesdiscussed above, the product of the PTEN gene is in-volved not only in migration but also in proliferation andsurvival (107).

F. Invasion Genes and Metastasis Genes,Separate Classes?

The multistep invasion process of metastasis ex-plains that invasion is a prerequisite for metastasis; itdoes, however, not account for the large differences inmetastatic ability of invasive tumors. A working definitionof metastasis genes, different from invasion genes, mightbe that their activation or inactivation changes the meta-static phenotype of invasive tumors. In the above-men-tioned example of crosses of Rip1Tag mice with miceexpressing dominant negative E-cadherin (Rip1dnE-cad),

no distinction can be made between the acquisition ofinvasion and metastasis. MTS1 (metastasin 1) possiblymeets the criteria of a metastasis promoter gene becauseit conveys metastatic capability upon invasive nonmeta-static tumors (4). Mice of the GRS/A strain carry a mousemammary tumor virus (MMTV) provirus and have a highincidence of mammary tumors, due to the proviral acti-vation of the oncogenes WNT and INT-2. Histologically,these tumors represent moderately differentiated invasiveadenocarcinomas. The GRS/A mice were crossed withtransgenic mice expressing the MTS1 gene in the lactatingmammary gland under the control of an MMTV promoter.These MTS1 transgenics do not develop mammary tu-mors. Successful crosses between GRS/A and MTS1 micehave mammary cancers that are not only locally invasivebut also form metastases in the lungs. Here, MTS1 acts asa metastasis-promoter gene, but the phenotypic alterationresponsible for the formation of metastases from invasiveprimary cancers is not clear.

FIG. 4. Genetic manipulation of apparentlyspontaneously immortalized cell lines from dog re-nal (MDCK) and mouse mammary (NMuMG) originresulting in E-cadherin-dependent alterations of in-vasion in vitro. [Schematic representation of datafrom Vleminckx et al. (432).]



on July 8, 2014

Dow

nloaded from

G. Noncancer Invasion-Suppressorand Invasion-Promoter Genes

During normal embryonic development, spatiotem-poral activation and inactivation of invasion-promotergenes and invasion-suppressor genes participate at theregulation of gastrulation and morphogenesis. The idea toconsider some of the proteins encoded by these genes aspromoters or suppressors of cancer invasion came fromembryology (390). For example, E-cadherin is first ex-pressed at the morula stage, hence its former name uvo-moruline, where it serves compaction, the earliest form ofepithelial organization (178, 224). At the onset of gastru-lation, when cells start to migrate from the ectodermundergoing an epithelial to mesenchymal conversion, E-cadherin is downregulated and N-cadherin is expressed.This switch of cadherin expression from DE- to DN-type(D for Drosophila) occurs downstream of the invasionpromoter genes Twist and Snail (302). Twist encodes anuclear protein containing a helix-loop-helix motif, whichprobably acts as a transcription factor. At gastrulation inDrosophila melanogaster, Twist-positive cells roll intothe presumptive mesoderm, as beautifully illustrated inReference 226; Twist / mutants, like Snail / mu-tants, fail to complete gastrulation (68). Morphogeneticactivities that act through activation of the E-cadheringene are found also in hepatocytes, through the CDH1-binding transcription factor hepatocyte nuclear factor(HNF)-4 (375) and in thyrocytes, through thyroid stimu-lating hormone (52). Reversion of an invasion-associatedphenotype, namely, mesenchymal to epithelial transition,is observed during metanephrogenesis. This transitionwas mimicked in the human fetal kidney cell line HEK293where expression of PAX-2, a member of the paired-boxhomeotic gene family, was associated with a gain of E-cadherin and -catenin expression (408).

In microorganisms, virulence genes are regulators ofinvasion. Historically, the first transformation from non-virulent into virulent Streptococcus pneumoniae formedthe basis for the identification of DNA as the geneticmaterial by Avery et al. (12). Transfection was applied tothe analysis of specific invasion genes first in bacteria(181). Transfer of a single genetic locus from the invasivebacterium Yersinia pseudotuberculosis made the nonin-vasive Escherichia coli invasive into cultured vertebratecells. When the noninvasive L. innocua is transfectedwith a plasmid harboring the internalin A (inlA) gene, thebacterium becomes invasive into Caco-2 cells (141). Theinvasion assay took advantage of the fact that invaded,hence intracellular, bacteria are protected against antibi-otics that fail to penetrate into the cells (see Fig. 3). InListeria, virulence genes regulate not only entry into thevertebrate cell, but also intracellular multiplication andspreading. Six of these virulence genes are clustered onthe bacterial chromosome (prfA, plcA, hly, mpl, actA, and

plcB), the two others (inlA and inlB) form a distinctoperon. All these genes are coordinately regulated byPrfA, the trancriptional activator encoded by the prfAgene (268). Such a gene, switching on and off coordinatedinvasion programs, has not been found in cancer cells, sofar. The invasive, virulent, and pathogenic E. histolyticadiffers genetically from the noninvasive, avirulent, andnonpathogenic E. dispar as evidenced by restriction frag-ment length polymorphism and sequencing of single copygenes (320). Here, like in bacteria, virulence genes arecrucial for invasion; downregulation of their expressionby antisense RNA transfections causes a reduction ofinvasion-associated molecules such as amoebapore (54),the 35-kDa light subunit of the Gal/GalNac specific lectin(8) and the cysteine proteinase 5 (9). The life cycle of theparasite Schistosoma mansoni provides an example ofspatiotemporal regulation of an invasion-promoter geneat the cercaria stage. Before the cercaria leaves its snailhost to swim freely in the water, the gene encoding aserine protease is switched on. This powerful lytic en-zyme is activated only upon contact with the human skinand in this way acts exclusively at the site of invasion.Once the cercaria has invaded into the dermis of its newhost, the protease gene is switched off (132).

Taken together, the above-mentioned examples ofinvasion-suppressor and invasion-promoter genes clearlydemonstrate that multiple genes are implicated in inva-sion and that they vary between different types of cells.Their activation or inactivation triggers programs of cel-lular activities that are rarely restricted to invasion andusually changes also the other cellular activities depictedfor cancer cells in Figure 1.

IV. CANCER CELLS, HOST CELLS, AND TUMORCELLS: ALL INVADERS

A. Cancer Cells and Host Cells

Cellular behavior and gene activation or inactivationare greatly influenced by the environment in normal aswell as in pathological situations including cancer (seeFig. 1). The Lancets first volume (313) launched the seedand soil hypothesis asking the question: What is it thatdecides what organ shall suffer in a case of disseminatedcancer? His answer is still valid: The microenvironmentof each organ (the soil) influences the survival and growthof tumor cells (the seed). Pathologists have since a longtime recognized that tumors contain not only neoplasticcells, further called cancer cells, but also host cells. Thehost participation at the establishment of the tumor isdescribed as desmoplasia, consisting of fibroblastic cellsand extracellular matrix, as inflammation and immuneresponse represented by lymphocytes, macrophages, anddendritic cells, and as angiogenesis evidenced by newly

CANCER INVASION 349


on July 8, 2014

Dow

nloaded from

formed blood and lymph vessels. These host elements,although more abundant in some types of cancer than inothers, are omnipresent. For example, less than half ofmost pancreatic cancers are occupied by cancer cells, themajority being host cells. In line with this histologicalobservation is the detection of a cluster of invasion-spe-cific expression of genes encoding molecules that partic-ipate at the reaction of the host (345).

There are clinical and experimental data to believethat host cells play a major role in invasion and metasta-sis. Metastasis may depend on the specific site of theprimary cancer. The frequency of distant metastasis fromsquamous cell carcinomas of the head and neck regiondepends on the subsite of the primary cancer, varyingfrom 3.1% for tumors situated in the larynx to 28.1% fortumors in the nasopharynx as summarized in Reference249. More recently, the latter figure has increased to40%as new treatment techniques such as intensity modulatedradiotherapy (IMRT) have improved local tumor controland patient survival (388). Orthotopic, compared withparatopic (usually indicating subcutaneous), implantationinto immunosuppressed mice provides an experimentaldemonstration of site specificty of invasion and metasta-sis. When human colon cancer cells are implanted in thewall of the cecum (325) or when oral squamous cellcarcinoma cells are implanted into the tongue of nudemice (202), they invade and metastasize in contrast totheir subcutaneous counterparts. Organ specificity indi-cates that some tumors metastasize more frequently tospecific organs than could be expected from their trans-port in the circulation and their passage through capillarynetworks. Examples of preferentially affected organs arethe brain for lung cancer and melanoma and bone forprostate and breast cancers (436). It is still a matter ofdebate whether organ specificity of metastasis is due tospecific homing and extravasation or to specific survivaland growth of the cancer cells at the site of extravasation.The type of host cells, e.g., endothelial cells, that partici-pate at tumor development all are invasive themselvesand some, e.g., leukocytes, are even metastatic. It is,therefore, justified to ask the question who is invadingwho (238). The recruitment of host cells is most likely theresult of the production by the tumor microecosystem ofstimulatory and inhibitory factors. Moreover, these hostcells may proliferate in the tumor ecosystem, again gov-erned by balances between inhibitory and stimulatoryfactors. Cancer cells may also cause transdifferentiationof host cells, e.g., fibroblasts into myofibroblasts.

B. Myofibroblasts: Stimulators of Invasion

The role of myofibroblasts, first described as smoothmuscle-like fibroblasts by Gabbiani et al. (140), in cancerinvasion has been recently reviewed (106). The emphasis

of this review is on the continuous molecular cross-talkbetween the cancer cells and the host (Fig. 5). Cytokines,such as TGF- and platelet-derived growth factor(PDGF), are released from the cancer cells, probably at aproinvasive state of tumor development; they stimulatethe transition of fibroblasts into myofibroblasts. The lattercells are found, indeed, more frequently in preinvasivelesions of the colon such as villous adenomas and FAP,that have a higher risk of transition into invasive carci-noma, than in lesions such as tubular adenomas, that havea lower risk of progression (260). Myofibroblasts do par-ticipate at numerous noncancerous pathological andphysiological processes. During wound healing they assistat migration, proliferation, and contraction. When thewound is closed, myofibroblasts undergo apoptosis, quitein contrast to tumors where they persist as in a woundthat does not close (115). This idea illustrates that cancercells operate by noncancer specific activities, but they failto regulate these activities properly. Myofibroblasts pro-duce numerous molecules, growth and motility factors,angiogenic factors, extracellular matrix components, andproteinases, that all promote the invasion and also thegrowth of cancer cells. Other molecules of putative inter-est for invasion expressed by myofibroblasts include-smooth muscle actin, vimentin, c-MET (404), proteo-lytic FAP displaying also dipeptidyl peptidase activity(316), cyclooxygenases (COX)-1 and -2 and N-cadherin,associated with -catenin, -catenin, p120CTN, and T-catenin (187, 425). An early demonstration of the inva-sion-stimulating activity of myofibroblasts resulted fromthe differential behavior in vitro compared with in vivo ofPROb cancer cells that were isolated from a chemicallyinduced rat colon tumor (108). PROb cells invaded nei-ther into collagen nor into Matrigel nor into embryonicchick heart in culture. Upon subcutaneous injection intosyngeneic rats, PROb cells did, however, produce invasivecancers, and numerous myofibroblasts were present atthe front of invasion in line with observations on humancancers. PROb cells were stimulated to invade also invitro, provided myofibroblasts were added to the culturesystem. These myofibroblasts themselves were also inva-sive. One interesting example of noncancerous myofibro-blast invasion is described in experimental tubulointersti-tial fibrosis (288, 386). There is good histological andultrasctructural evidence to accept that tubular epithelialcells transdifferentiate into myofibroblasts invadingthrough a disrupted basement membrane into the under-lying stroma. Like in cancer, TGF- produced by thetubular epithelial cells stimulates fibrosis (322).

C. Angiogenesis Before Invasion

Blood vessels and lymph vessels provide tumors withnutrients and cytokines, necessary for growth and inva-



on July 8, 2014

Dow

nloaded from

sion; they provide the routes for systemic spread of can-cer cells; and they mediate the communication betweenthe primary tumor and its metastasis (309). These vesselsrepresent the response of existing blood and lymph ves-sels to balances between positive and negative angiogenicfactors produced by the cancer cells. The type of vessels,hematogenic or lymphatic, that are invading the tumormight be determined by the type of vascular endothelialgrowth factor (VEGF) produced (368). Because invasioninto vessels initiates metastasis, the type of VEGF mightalso determine the route of metastasis, lymphogenic orhematogenic. In exceptional cases, the cancer cells them-selves may transdifferentiate into endothelioid cells andform the tumor vascular system, a phenomenon that iscalled vasculogenic mimicry (244). Several excellent re-views on tumor angiogenesis were produced (70, 134,207). We would, therefore, limit the discussion to theobservation that neoangiogenesis may be needed for pri-mary invasion as well as it is for growth as evidenced byan in vivo mouse model (14, 15). Transplantation of col-lagen gels coated with malignant murine keratinocytes onthe dorsal muscle fascia of wild-type mice resulted ininvasive squamous cell carcinomas. When plasminogenactivator inhibitor (PAI)-1 / knock-out mice wereused, the cancer cells failed to invade, and there was noangiogenesis. Intravenous injection of a recombinant ad-

enovirus vector carrying the human PAI-1 cDNA restoredangiogenesis and invasion. The molecular explanation isthat plasmin proteolysis must be tightly controlled toallow vessel stabilization and differentiation.

D. Tumor Infiltrated Leukocytes:Helpers of Invasion

Tumor tissues are frequently infiltrated by host leu-kocytes, sent in by the immune system of the host in anattempt to reject the tumor. Indeed, some of these hostcells are able to kill cancer cells or to secrete antiangio-genic factors. A recent example is provided by the highsusceptibility to skin carcinogenesis of mice lacking Tcells (149). It is, however, evident that such infiltratedcells can also have tumor-promoting effects. This may beillustrated by the earlier finding that nonmetastatic lym-phoma cells become metastatic upon fusion with acti-vated leukocytes (96). The countercurrent principle orig-inally developed in chemistry to separate oil lovers fromwater lovers and, later, in Drosophila genetics of behaviorto separate light lovers from dark lovers, was applied tothe helper function of leukocytes in invasion (308). Can-cer cells produce chemotactic cytokines (called chemo-kines), a family of small proteins that attract leukocytes

FIG. 5. Micro-ecosystem of a primary cancer.Cellular activities associated with invasion impli-cate reduced cell-cell adhesion, altered cell-matrixadhesion, migration, ectopic survival, and lysis ofextracellular matrix. Most of these activities aremodulated by the cross-talk between cancer cellsand host cells, which are activated by cytokinesreleased from the cancer cells. [Adapted from Ma-reel et al. (251) and De Wever and Mareel (106).]

CANCER INVASION 351


on July 8, 2014

Dow

nloaded from

from the circulation along a chemical gradient toward thetumor. These chemokines also stimulate the productionof MMPs by the attracted leukocytes, that dissolve theextracellular matrix (ECM) on their way to the tumor. Bydoing so, a tunnel is created for the invading cancer cells(Fig. 6). Moreover, the chemokines act as growth factorsfor the cancer cells and are angiogenic, so providingvessels that mediate invasion and serve as routes formetastasis.

E. Osteoclasts: Targets for Therapy

In bone metastasis, a frequent complication in malig-nant tumors, cancer cells subvert the host dynamic ho-meostatic mechanisms that preserve the structure andfunction of the skeleton. The result is excessive break-down of bone, pain, and eventual fractures. Cancer cellsrelease osteoclast activating factors, such as interleu-kin-1, tumor necrosis factor, TGF-, epidermal growthfactor (EGF), PDGF, and prostaglandins, and activatedosteoclasts cause breakdown of bone matrix (Fig. 7). Thisbreakdown is probably not limited to osteoclasts as it hasbeen attributed also to cancer cells and even to osteo-blasts and osteoblast-like cells (221). Bone matrix break-

down releases chemotactic factors attracting cancer cellsand growth factors stimulating their proliferation. Osteo-clasts are successful targets for treatment of bone metas-tasis by bisphosphonates acting through inhibition of os-teolytic activity, fortifying bone matrix and interferingwith the formation of osteoclasts from monocytes (172,282). Bisphosphonate acts not exclusively on osteoclastsas it induces also apoptosis in human breast cancer cellsin experimental animal models. The naturally occurringdecoy receptor osteoprotegerin, a member of the tumornecrosis receptor family, also inhibited metastatic osteol-ysis and decreased tumor burden in experimental models(278). Osteoprotegerin antagonizes the binding betweenthe osteoclast RANK receptor and its ligand, which isnecessary for osteoclast differentiation.

F. Molecular Cross-Talk inNoncancerous Situations

Epithelial-mesenchymal interactions are crucial inmorphogenesis (158, 367). In adults, invasive and meta-static normal leukocytic stem cells are retained in thebone marrow through adhesion to stromal cells. Microbialinvasion is to a large extent governed by host factors.Species specificity is a striking example of host-deter-mined invasion. L. monocytogenes invades human,chicken, rabbit, and guinea pig intestine but neithermouse nor rat, a phenomenon that is explained throughsingle amino acid differences in the first extracellulardomain of the enteric E-cadherin, serving as the receptorfor the bacterial virulence factor inlA. In amoebiasis,caused by E. histolytica, human species specificity isexplained by the transformation from noninvasive cyststo invasive trophozoites depending on a number of humanhost cell factors such as acid in the stomach, pancreaticenzymes, low oxygen content, inorganic salts, and micro-flora (148). Organ specificity of metastasis is illustrated inleishmaniasis. According to the type of leishmania, le-sions will develop at the site of injection in the skin (L.ropica) or in the mucocutaneous tissues of the respira-tory and the genital tract (L. brasiliensis) or in the vis-cera, mainly liver and spleen (L. donovani). Although theexact mechanisms of this organ specificity are unknown,immune mediators may be involved, since the cutaneousforms of leishmaniasis also affect the viscera in HIV-positive individuals. Examples of parasites that are trans-ported by host cells, a kind of passive invasion, areTrypanosoma and Leishmania using macrophages andPlasmodium merozoites using erythrocytes (200).

The genomic alterations in epithelial cells, whichlead to cancer, disturb the molecular conversation be-tween the epithelial cells and the underlying stroma. As aconsequence, the cancer cell population, developing atthe primary tumor site, attracts host cells and so creates

FIG. 6. Schematic of the countercurrent model of leukocyte-assistedcancer invasion, as proposed by Opdenakker and Van Damme (308).



on July 8, 2014

Dow

nloaded from

a dynamic tumor microecosystem in which there exists acontinuous cross-talk between the cancer cells and thehost cells. Such microecosystems are created at each stepof invasion by cancer cells and also by normal cells and bymicrobial organisms. The type of elements participatingat these microecosystems is somewhat different for eachof the invasive steps. A characteristic of an ecosystem isthat alteration of a single element may dramaticallychange the entire system. The transition from the nonin-vasive, i.e., maintenance of normal tissue architecture orabsence of invasion, to an invasive phenotype within themicroecosystem constitutes such a dramatic change thatmay equally depend on alterations in elements of the hostor in the potentially invader population (see Fig. 5).

V. CELLULAR ACTIVITIES ASSOCIATED WITHTHE INVASIVE PHENOTYPE

Cellular activities positively or negatively associatedwith the invasive phenotype comprise cell-cell adhesion,cell-matrix adhesion and ectopic survival, migration, andproteolysis (see Fig. 5). In cells that have progressedtoward malignancy through activation of promoter genesand inactivation of suppressor genes (see Fig. 1), thesecellular activities are regulated by autocrine and para-crine ligands, resulting in modulation of the invasive phe-notype. The challenge is to trace the pathways from theligand to the receptor to the signal transduction and fi-nally to the cellular response that is crucial for the alter-

FIG. 7. Schematic of the micro-ecosystem ofbone metastasis. Bisphosphonates inhibit os-teoclastic bone destruction and so counteract thefurther development of bone metastasis. E-cad-herin is implicated in the fusion of osteoclast pre-cursor monocytes. [Modified from Mareel et al.(249).]

CANCER INVASION 353


on July 8, 2014

Dow

nloaded from

ation of the invasive phenotype. Methods used implicatethe genetic manipulation of cells to change their sensitiv-ity toward putative invasion modulators and the use ofpharmacological inhibitors targeting specific elements ofthe signaling pathways. Such signaling pathways can bepresented in a linear fashion, connecting the initiatingevent, e.g., activation of a receptor tyrosine kinase byligand binding, to the activation of gene expression in thenucleus. However, signal transduction pathways havemany potential branch points that provide opportunitiesfor signaling cross-talks, and most invasion pathways arecomplex with multiple branch points rather than simplylinear (113). In the latter scenario, it is expected that themost highly connected proteins are the most importantones for the cells most vital functions (190). The numberof molecules and networks implicated in invasion-associ-ated cellular activities has increased to the point that acomprehensive review falls beyond the scope of thepresent review. As announced earlier, the following dis-cussion deals with selected examples.

A. Cell-Cell Adhesion

Homotypic (between cells of the same type) epithe-lial cell-cell adhesion counteracts the escape of cells intoneighboring tissues and, even, invasion by well-differen-tiated epithelial strands implicates some degree of weak-ening of cell-cell adhesion, as demonstrated in a directway for the first time by Coman (80). Heterotypic cell-celladhesion, e.g., between circulating cancer cells and thevascular endothelium, promotes invasion through the vas-cular wall and the initiation of metastases. On rare occa-sions, homotypic cell-cell adhesion may promote inva-sion, mediating apical implantation of cancer cells intothe bladder mucosa (331) or lymphovascular permeationin inflammatory breast cancer (407).

1. Homotypic cell-cell adhesion: theE-cadherin/catenin/actin pathway

The prototype homotypic epithelial cell-cell adhesionmolecule operating through homophyllic interaction isE-cadherin. Adherens junctions are the structure in whichE-cadherin operates. Other intercellular junctions, men-tioned as putative players in invasion, are tight junctions,desmosomes, and gap junctions. We have provided evi-dence (see sect. III) in favor of an invasion-suppressorfunction of E-cadherin. The question is whether or nothomotypic cell-cell adhesion is responsible for suppres-sion of invasion. E-cadherin is certainly suited for suchfunction as dimers on one cell form stable molecularbonds with E-cadherin dimers on another cell. Such ex-tracellular interaction is sufficient for a relatively weakadhesion, whereas stronger adhesion necessitates alsointracytoplasmic interactions of E-cadherin (60) and an

intact transmembrane domain (177). In many experi-ments a positive correlation was found between lack ofinvasion and cell-cell adhesion in vitro, the E-cadherinspecificity of both phenotypes being demonstrated withthe aid of functionally inactivating antibodies. (30, 46, 81,253, 296, 390, 423). The expression of large proteoglycanssuch as episialin, also called MUC-1, may sterically hinderthe homophilic interactions of E-cadherin and so contrib-ute to invasion of cancer cells (380, 433). In vivo,Rip1dnE-cad mice, overexpressing a dominant negativeform of E-cadherin in pancreatic -cells (see sect. III),showed a transient perturbation of -cell aggregation dur-ing islet development (91). Other homotypic cell-cell ad-hesion complexes have also been implicated in invasion.Desmosomes are considered as regulators of epithelialstructure and as invasion suppressors by some authors,and the experimental arguments resemble the ones putforward for adherens junctions (342, 413). These authorspresume that desmosomal cadherins and E-cadherin arecomparably involved in the maintenance of the epithelialstructure. L929 fibroblasts transfected with desmosomalcomponents, namely, desmocollin, desmoglein, and plako-globin, become adhesive and noninvasive, and these re-verted phenotypes are lost upon addition of short peptidescorresponding to cell-cell adhesion recognition (CAR) sitesof desmosomal cadherins. Furthermore, downregulation ofdesmosomal adhesion molecules correlates with invasionand metastasis in some human cancers without consider-ation of the E-cadherin status (95, 287). In invasive andmetastatic cancers of the oral cavity, downregulation ofdesmosomal cadherins was associated with loss of E-cadherin (364). Gap junctions are membrane-spanningchannels composed of protein subunits called connexins;they facilitate intercellular communication through small(Mr 1,000) molecules. Such junctions may act as invasionpromoters, since transfection with connexins and homo-typic coupling lead HeLa cells to invade into chick heartfragments in organ culture (156). In human breast cancercell lines, relative amounts of different connexins wereheld responsible for metastasis (350).

Although most attention has been paid to the cell-celladhesion function of E-cadherin, it can certainly not beexcluded that other invasion-associated cellular activitiesare implicated. Loss of E-cadherin, eventually associatedwith upregulation of N-cadherin, may stimulate migrationand so promote invasion (55, 163). In cancer of the pros-tate (406, 409) and of the head and neck mucosa (182),migration of epithelial cells coincides with the loss ofE-cadherin and is sometimes accompanied by a gain ofN-cadherin expression. The E-cadherin/catenin complexis implicated also in directed migration during epithelialwound healing, presumably through contact inhibition ofmembrane ruffling (6, 81) and in intercellular motilityduring epithelial morphogenesis (158). This is in line withthe observation that a monoclonal antibody against E-



on July 8, 2014

Dow

nloaded from

cadherin neutralizes the invasion-stimulating effect of bileacids on colon cancer cells in vitro (97). Finally, thereexists good evidence to accept that E-cadherin downregu-lation stimulates growth and ectopic survival (197, 378).

The molecular organization and the links of the E-cadherin/catenin complex with multiple other receptorand nonreceptor signal-tranducing systems make it verylikely that alterations of an element of the complex affectsmultiple cellular activities (Fig. 8). Next to the genetic andtranscriptional regulation discussed in section III, the E-cadherin/catenin complex can be structurally and func-tionally downregulated at the posttranscriptional levels inmany ways, including phosphorylation (323), glycosyla-tion (457), proteolysis (295), endocytosis (47), and steri-cal hindrance (171, 380, 433). Consequently, there existmany natural and pharmacological agents that stimulateor inhibit invasion of cells by direct or indirect action onthe E-cadherin/catenin complex (See Table 3 in Ref. 419).Armadillo repeats, as present in -catenin, plakoglobin,and p120CTN, are good substrates for tyrosine kinases(93). Indeed, tyrosine phosphorylation of -catenin bynonreceptor kinases, like Src (31), and receptor tyrosinekinases like EGFR (365) and c-Met (363), perturbs the

E-cadherin/catenin complex in various manners: alter-ations of the binding of the complex to actin; disturbanceof the signaling function through conformational changesthat hamper the cross-talk with other proteins; and re-cruitment of new partner proteins that possess phospho-tyrosine-specific binding domains. Abrogation of EGFR-induced tyrosine phosphorylation of -catenin inhibitsinvasion of melanoma cells (44). Furthermore, the -cate-nin binding site of E-cadherin shows Pro-Glu-Ser-Thr(PEST) sequences, candidates for degradation by theubiquitine/proteasome pathway, so that binding of -cate-nin prevents the cytoplasmic tail of E-cadherin from deg-radation (176). -Catenin links the E-cadherin/catenincomplex to the actin cytoskeleton (337). Successful trans-fection of -catenin-negative adhesion-deficient variantsof the HCT-8, coined HCT-8/R, colon cancer cell line withT- or E-catenins restored cell-cell adhesion and com-paction (187). p120 catenin is encoded by CTNND1 on11q11 (205). In cancer cells, p120CTN can be found at themembrane and in the nucleus, where it binds to the zincfinger transcription factor Kaiso (94, 424). Unlike -cate-nin, p120CTN is not degraded through interaction with theAPC/GSK-3 complex. Larger aggregates were formed by

FIG. 8. The E-cadherin/catenin and the wnt/-catenin signaling pathways. For detailed description, see sections IIIand V. Definitions are as follows: APC, adenomatous polyposis coli protein; Cdc42, cell division cycle 42, a smallG protein; CTN, catenin; DSH, dishevelled; EB1, microtubule end binding protein1; E-CAD, E-cadherin; Frz, frizzled; GF,growth factor; GSK-3, glycogen synthese kinase-3b; ILK, integrin-linked (serine/threonine) kinase; IQGAP1, GTPase-activating protein; Kaiso, transcription factor of the Pox virus and zinc finger superfamily; LEF, leukocyte enhancerfactor; MMP-7 (matrilysin), matrix metalloproteinase-7; PI3-K, phosphoinositide 3-kinase; PKB, protein kinase B; PTP,protein tyrosine phosphatase ; Rac, Ras-related C3 botulinum toxin substrate; Ras, small GTPase of the Ras homologfamily; Rho, Ras homolog; RTK, receptor tyrosine kinase; Shp-1, SH2 domain protein tyrosine phosphatase; Src, tyrosineprotein kinase encoded by the cellular protooncogene homologuous to the viral Rous sarcoma oncogene v-SRC; Smad3,Mad (mothers of decapentaplegic)-related protein 3, recognizing a palindromic DNA sequence; TFF, trefoil factors; Wnt,vertebrate homolog of Wingless. [Adapted from Van Aken et al. (419), with data from McCartney and Peifer (266), Mulleret al. (281), Lickert et al. (235), Kuroda et al. (219), and Wu et al. (451).]

CANCER INVASION 355


on July 8, 2014

Dow

nloaded from

cells that were transfected with wild-type E-cadherincompared with E-cadherin mutants, to which p120CTN

failed to bind (403). Current views hold that p120CTN

induces E-cadherin clustering and cell-cell adhesion bybinding to the juxtamembrane part of E-cadherin and thatintracellular signaling can induce an inactive form ofp120CTN (312, 403). Tyrosine phosphorylation of p120CTN

increases its affinity for the juxtamembrane domain oftype I cadherins and lowers its affinity for RhoA. RhoA,uncoupled from 120CTN, could subsequently be activatedby GEFs guanine nucleotide-exchange factors (GEFs)with further downstream RhoA sig

Documents

Aspectos celulares y moleculares del cáncer