Tissue Invasion and Metastasis - Molecular, Biological and Clinical - Review

Please citeCancer Bi

ARTICLE IN PRESSG ModelYSCBI-1183; No. of Pages 32Seminars in Cancer Biology xxx (2015) xxxxxx

Contents lists available at ScienceDirect

Seminars in Cancer Biology

j o ur na l ho me page: www.elsev ier .com/ locate /semcancer

Review

Tissue invasion and metastasis: Molecular, biological and clinicalperspectives

W.G. JianS.K. ThomK. HonokL. Yea, WS. Chens,F. Pantana Cardiff Univerb National Cancc University Hod University of e Royal Adelaidf Department og CEINGE Bioteh University of i Purdue Researj University of Mk Nara Medicall Physics Deparm University ofn United Arab Eo University of p Creighton Unq SASTRA Univer University of s Ovarian and Pt Wayne State Uu University of v New York Mew Mayo Clinic Cx University Ca

a r t i c l

Keywords:Cancer metastInvasionCancer therap

CorresponCardiff Univer

E-mail add

http://dx.doi.o1044-579X/ this article in press as: Jiang WG, et al. Tissue invasion and metastasis: Molecular, biological and clinical perspectives. Seminol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

ga,, A.J. Sandersa, M. Katohb, H. Ungefrorenc, F. Gieselerc, M. Princed,psone, M. Zollo f,g, D. Spanog, P. Dhawanh, D. Sliva i, P.R. Subbarayanj, M. Sarkar j,i k, H. Fujii k, A.G. Georgakilas l, A. Amedeim, E. Niccolaim, A. Aminn, S.S. Ashrafn,.G. Helfericho, X. Yango, C.S. Boosanip, G. Guhaq, M.R. Ciriolor, K. Aquilanor,

A.S. Azmit, W.N. Keithu, A. Bilslandu, D. Bhaktaq, D. Halickav, S. Nowsheenw,ox, D. Santinix

sity, Cardiff, United Kingdomer Center, Tokyo, Japanspital Schleswig-Holstein, Lbeck, GermanyMichigan, Ann Arbor, MI, USAe Hospital, Adelaide, Australiaf Molecular Medicine and Medical Biotechnology (DMMBM), University of Naples Federico II, Naples, Italycnologie Avanzate, Naples, ItalyNebraska Medical Center, Omaha, USAch Park, Indianapolis, IN, USAiami, Miami, FL, USA

University, Kashihara, Japantment, School of Applied Mathematical and Physical Sciences, National Technical University of Athens (NTUA), Athens, Greece

Florence, Florence, Italymirates University, Al Ain, United Arab Emirates and Faculty of Science, Cairo University, EgyptIllinois at Urbana-Champaign, Urbana, IL, USAiversity, Omaha, NE, USArsity, Thanjavur, India

Rome Tor Vergata, Rome, Italyrostate Cancer Research Trust Laboratory, Surrey, United Kingdomniversity, Detroit, MI, USA

Glasgow, Glasgow, United Kingdomdical College, Valhalla, NY, USAollege of Medicine, Rochester, MN, USAmpus Bio-Medico, Rome, Italy

e i n f o

asis

y

a b s t r a c t

Cancer is a key health issue across the world, causing substantial patient morbidity and mortality. Patientprognosis is tightly linked with metastatic dissemination of the disease to distant sites, with metastaticdiseases accounting for a vast percentage of cancer patient mortality. While advances in this area havebeen made, the process of cancer metastasis and the factors governing cancer spread and establishment atsecondary locations is still poorly understood. The current article summarizes recent progress in this areaof research, both in the understanding of the underlying biological processes and in the therapeutic strate-gies for the management of metastasis. This review lists the disruption of E-cadherin and tight junctions,key signaling pathways, including urokinase type plasminogen activator (uPA), phosphatidylinositol 3-kinase/v-akt murine thymoma viral oncogene (PI3K/AKT), focal adhesion kinase (FAK), -catenin/zincnger E-box binding homeobox 1 (ZEB-1) and transforming growth factor beta (TGF-), together withinactivation of activator protein-1 (AP-1) and suppression of matrix metalloproteinase-9 (MMP-9) activ-ity as key targets and the use of phytochemicals, or natural products, such as those from Agaricus blazei,Albatrellus conuens, Cordyceps militaris, Ganoderma lucidum, Poria cocos and Silybum marianum, together

ding author at: Cardiff-Peking Cancer Institute and Cardiff-Capital Medical University Joint Centre for Biomedical Research, Cardiff University School of Medicine,sity, Henry Wellcome Building, Heath Park, Cardiff CF14 4XN, United Kingdom. Tel.: +44 29 20687065.ress: [email protected] (W.G. Jiang).

rg/10.1016/j.semcancer.2015.03.0082015 Published by Elsevier Ltd.


ARTICLE IN PRESSG ModelYSCBI-1183; No. of Pages 322 W.G. Jiang et al. / Seminars in Cancer Biology xxx (2015) xxxxxx

with diet derived fatty acids gamma linolenic acid (GLA) and eicosapentanoic acid (EPA) and inhibitorycompounds as useful approaches to target tissue invasion and metastasis as well as other hallmark areasof cancer. Together, these strategies could represent new, inexpensive, low toxicity strategies to aid inthe management of cancer metastasis as well as having holistic effects against other cancer hallmarks.

1. Introdu

The chaicells, whethplex. Maligngrowth potmetastasizesome extentasis may vwithin the as those wi

Despite prevention visible usintime they cearly and lametastasis.for over 90[1,4,5]. Despison to othepartly due tto a lack of However, shas been wbe elucidateand a numbtruly under

Diagnosthe constantreatmentsstill associathe complein Fig. 1) antumors, fulthe discovemajor missto discuss kmetastasis small molecstrategies fo

2. Cellular

2.1. Cellce

In canceand cellceent primarythrough muof the primtion and m[79]. Key between cetight junctifer a weak ajunctions plial cells. Peof adhesion

in fas, plvolve7kip2

l epit due, red

met conterin lial

2.4)ffer ougave regay thehomohese smo

ed tonclule ef-EFA

Claud TJ cf epicirclrve tepar) thethe ptissuant i

ande intlD3,

(curtoplas, an

embrh efs a nteinsasis urinnd trted w

C (PKdies onrmpoction

n of events leading to the malignant transformation ofer through genetic or epigenetic alterations, is com-ant cells possess key hallmarks, namely, uncontrolled

entials and the ability to invade surrounding tissues and [1]. Cancer cells likely possess these innate abilities tot, though the degree and timing of invasion and metas-ary due to the genetic and epigenetic heterogeneitytumor and further signals from extrinsic factors, suchthin that particular microenvironment [2].substantial effort dedicated to the early detection andof cancer, most patients are likely to have micro- (notg conventional methods) or macro- metastases by theome to medical attention [3,4]. Cancer patients, bothte stage, dependent on life span, are likely to develop

This metastatic spread of the primary tumor accounts% of patient mortality associated with solid cancersite this, research into the eld of metastasis, in compar-r key events such as proliferation, etc., is lagging. This iso the complexity of the metastatic process but also duesufcient funding and efforts into this area of research.ignicant progress in this vital area of cancer researchitnessed over the past decade, though much remains tod before we fully understand this pernicious conditioner of signicant gaps remain to be lled before we canstand this complex process.is and treatment of metastatic disease are vital areas int battle many patients face against cancer, yet effective

are limited and substantial morbidity and mortality areted with metastatic disease [5,6]. This, together withxities surrounding the metastatic process (summarizedd the complex nature and heterogeneity of metastaticly supports and justies further research dedicated tory of a less toxic means to treat this condition. This is theion of getting to know cancer (GTKC). This review aimsey knowledge gaps, explore potential targets in tacklingand also potential methods, including phytochemicals,ule inhibitors and natural compounds in devising newr treating metastasis.

properties and metastasis

ll adhesion

rs derived from the epithelium, inter-cellular structuresll adhesion are key factors in maintaining a coher-

tumor mass [7,8]. Abnormalities in these structures,tation or dysregulation, can lead to the dissociationary tumor and an enhanced potential for dissemina-etastatic spread of cancer cells to secondary locationsstructures involved in maintaining this adhesivenesslls include adherens junctions (including desmosomes),ons (TJ) and gap junctions. While gap junctions con-

cadhercateninalso inand p5normamainlyHencetial forloss ofE-cadhepitheSectioncould otion, ththere hin this nameland didiet. Tand dereporttypes idesirabby non[16].

2.1.1. The

types othat enThey seically sand (2along minal importbarrierinto thmarvefamilyThe cyproteinthe mresearcthere iTJ prometastof TJs dtion, aassociakinase

Stunow cgral co this article in press as: Jiang WG, et al. Tissue invasion and metastasis:ol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

dhesion structure and TJs, a modest one, the adherensrovide the most powerful source of adhesion in epithe-rhaps one of the strongest and most studied regulators

is E-cadherin (cadherin-1 or CDH1), a member of the

the cellulartial, howeva series of pdevelopme 2015 Published by Elsevier Ltd.

mily of proteins. E-cadherin, together with associatedays a key role in maintaining cellcell adhesion and isd in the regulation of the cell cycle regulators p27kip1

, which are involved in cellcell contact inhibition inhelium, but which are lost or disturbed in cancer cells,

to the loss of E-cadherin in cancer cells [8,10,11].uced cellcell adhesion not only enhances the poten-astatic dissemination of cancer cells but also, throughact inhibition, promotes uncontrolled cell growth [7].has also been established as a key mediator of the

mesenchymal transition (EMT) process (discussed in. Thus, enhanced expression of key cadherin moleculespotential as a strategy to control metastatic dissemina-h realizing this potential has proved difcult; thus farbeen few reports identifying viable treatment optionsrd. However, there are a number of noteworthy options,

polyunsaturated fatty acids gamma linolenic acid (GLA)--linolenic acid (DGLA), both obtainable through thehave been reported to be key regulators of E-cadherinsomal cadherins in cancer cells and have also been

have benecial effects for patients with several cancerding pancreatic cancer and breast cancer [1215]. Thefects of these essential fatty acids (EFAs) were blocked, as long chained oleic derivatives on human cell lines

ins in canceromplex is the apical most junctional complex in mostthelial and endothelial cells. TJs are the gasket-like sealse each columnar epithelial cell around its apical pole.wo roles: (1) they help to maintain cell polarity by phys-ating the apical and the basolateral membrane domainsy prevent free interchange of substances by diffusionaracellular pathway between the luminal and antilu-e uid compartments. TJs and their permeability aren the formation of the blood brain barrier, blood retinal

blood testis barrier. The TJ proteins can be sub-dividedegral membrane proteins such as occludin, tricellulin,junctional adhesion molecules (JAMs) and the claudinrently 27 members [17]) and the cytoplasmic proteins.smic adaptor proteins are the zonula occludens or ZOd are designated ZO-1, -2, and -3. These proteins linkane proteins to the actin cytoskeleton. Traditionally,forts focused on barrier and fence functions, however,ew movement in the eld, which is to understand how

participate in cell proliferation, transformation, andsuppression. Recent studies have demonstrated the roleg epithelial tissue remodeling, wound repair, inamma-ansformation into tumors. Epithelial multilayering wasith increased TJ permeability [18], activation of proteinC)- [19] and phosphorylation of TJ proteins [20].

focusing on the molecular architecture of the TJ havemed that the claudin family of proteins is the inte-nent of the TJ. Loss of cellcell adhesion is central to Molecular, biological and clinical perspectives. Semin

transformation and acquisition of metastatic poten-er, the role the claudin family of proteins may play inathophysiological events, including human carcinomant, is only now beginning to be understood. Several

Please cite this article in press as: Jiang WG, et al. Tissue invasion and metastasis: Molecular, biological and clinical perspectives. SeminCancer Biol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

ARTICLE IN PRESSG ModelYSCBI-1183; No. of Pages 32W.G. Jiang et al. / Seminars in Cancer Biology xxx (2015) xxxxxx 3

Fig. 1. The metastatic cascade and potential for therapeutic interruption. Changes in cellular properties are necessary to allow the development of an invasive phenotypeand progression through the metastatic cascade. Key events of the cascade are outlined. Targeting such properties/events or the underlying signaling pathways using lowtoxicity drugs holds great potential to disrupt cancer cell progression through this cascade.

Please cite asis:Cancer Bi


claudin mouse knockout models have demonstrated their impor-tant role in the maintenance of tissue integrity in various organs.The mechanisms of claudin regulation and their exact roles in nor-mal physiology and disease are being elucidated, but much workremains to

There ardivided in csimilarity [and 19 andClaudins arlial cells [2in cell memEL1 and ELacid long inthe intracelThe carbox(post synappressor, anand ZO-3 [2selectivity wing claudinfunctions aanisms [26claudin expin the EMT claudin mecal factors iand disruptcations, incof claudins the regulatiincreases anproteins hasites in theprotein kinato decreasefor claudin-be phosphoquently actassociated A1 (EphA1)complex wregulation thermore, msignaling, s1/2 and p38a profoundare also rembreaks and an importanby a uniquof the TJ ar[24]. Host fclaudin expclaudin endcytokines, s(IL)-13, dowparacellular

Growth tion of cell factor (EGFgrowth factlular distrib[28,29,41]. family [42]epithelial c

in vitro invasive behavior. Snail and Slug bind to the E-box motifspresent in the human claudin-1 promoter which play a critical neg-ative regulatory role in breast cancer cell lines that expressed lowlevels of claudin-1 [42]. Caudal type homeobox 2 (Cdx-2), hepa-

NuclA-4)

gen thatn tuors a,46].portafuncing wre inratios th

in bhowted h MAecre

[49]-3 a

ncrea [38]masexprrcinctabcervi

that-4, xpreer cessibiion. udinse to

canction oentiatme

the blarrela

al blailenc

carcistonenou

vely, leadasy utesnism

claure anso afed inyl telcellar tlocaludiegene this article in press as: Jiang WG, et al. Tissue invasion and metastol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

be done.e 27 types of claudins in mammals [17,21] and they arelassic and non-classic claudins based on their sequence21]. Classic claudins include claudins 110, 14, 15, 17

non classic claudins 1113, 16, 18 and 2024 [21].e found in epithelial, mesothelial, glial and endothe-224] with a molecular weight of around 20 kDa andbranes they are composed of two extracellular loops,2, four transmembrane domains, one small 20 aminotracellular part between the two extracellular loops andlular aminoterminal and carboxyterminal ends [21,25].yterminal end has regions which recognize the PDZtic density protein, Drosophila disk large tumor sup-d zonula occludens-1 protein) domains of ZO-1, ZO-25]. The larger EL1 loop inuences paracellular chargehereas the smaller EL2 loop binds to the correspond-

of the neighboring cell [25]. Claudin expression andre regulated at multiple levels and by diverse mech-,27]. An important question related to regulation ofression and cancer is the role that claudins may playprocess [28,29]. The paracellular barrier modulated bymbers can be affected by a wide range of physiologi-ncluding cell signaling pathways, hormones, cytokines,ion of the cellcell contacts. Post-translational modi-luding phosphorylation, lipid modication and removalby endocytosis, appear to be potential mechanisms foron of claudins. Phosphorylation has been linked to bothd decreases in TJ assembly and function. Most claudinve putative serine and/or threonine phosphorylationir cytoplasmic carboxyterminal domains. For instance,se A (PKA) mediated phosphorylation has been shown

assembly of claudin-3 into TJs [30], yet is necessary16 assembly and function [31]. Claudin-3 and -4 canrylated in ovarian cancer cells by PKA, a kinase fre-ivated in ovarian cancer [30]. Claudin phosphorylationwith TJ disassembly is also enhanced by EPH receptor, which is recruited to bind to claudin-4 by forming aith ephrin-B1 [32]. Studies have implicated PKC in theof TJs through phorbol ester stimulation [30,33]. Fur-odulation of mitogen-activated protein kinase (MAPK)

pecically extra cellular signal-regulated kinase (ERK), as well as phosphatidylinositol 3-kinase (PI3K) have

effect on TJ sealing and claudin expression [30]. TJsodeled at a more macroscopic level through strand

reformation [34]. Clathrin-mediated endocytosis playst role in this process [35,36]. Claudins are internalized

e mechanism, where the tightly opposed membranese endocytosed together into one of the adjoining cellsactors and cytokines can also inuence TJ turnover andression [37], for instance, interferon (IFN)- increasesocytosis and TJ permeability [38]. Other inammatoryuch as tumor-necrosis factor (TNF)- and interleukinn regulate claudins and induce a marked increase in

permeability by epithelial cells in culture [39,40].factor receptors that are important in the regula-proliferation and survival including epidermal growth), hepatocyte growth factor (HGF) and insulin likeor (IGF) receptors regulate claudin expression and cel-ution though once again in cell/tissue specic mannerClaudin transcription can be regulated by the Snail/Slug. It is well established that overexpression of Snail inells induces EMT and the acquisition of migratory and

tocyte4 (GATclaudinshowna knowinhibitels [45the imtissue regardsignatu

Altesis begtissueshave sassociathrougto be dcancerclaudining pacancercarcinolower tatic caundetein the shownclaudintheir eof cancthe poformaton claing duhumanregulawas idand tresion ofhumanalso cotumortially sbreastand hiendogeIntuitimight sia is econtribmechacertainstructumay alinvolvcarbox

Celin cellubrane best sthetero Molecular, biological and clinical perspectives. Semin

ear Factor 1-alpha (HNF-), and GATA binding protein [43,44] can bind to the promoter regions of variouses and affect their expression. Furthermore, it has been

colonic claudin-1 transcripts are regulated by Smad-4,mor suppressor as well as histone deacetylase (HDAC)nd thus support a complex regulation at multiple lev-

Collectively, the data provides an emerging picture ofnce of claudin homeostasis in normal and pathological

tion, but there remains much to be learned, especiallyhether it may be possible to identify a distinct claudin

the initiation and progression of various tumor types.ns in claudin expression proles during tumorigene-

e question of how claudins are regulated in differentoth normal and pathological situations. Tan et al. [47]n that the expression and distribution of claudin-1 iswith cell dissociation status in pancreatic cancer cellsPK 2 activation. By contrast, claudin-7 has been found

ased in invasive ductal carcinomas [48], head and neck and metastatic breast cancer [37]. On the other hand,nd -4 are frequently elevated in various cancers includ-tic ductal adenocarcinoma, prostate, uterine, ovarian

and breast cancer [50] while hepatocellular and renal expressed lower levels of claudins-4 and -5 [22]. Whileession of claudin-2 was also seen in breast and pros-omas, expressions of claudin-1 and claudin-7 that werele in normal cervical squamous epithelium increasedcal neoplasia [22,51]. Intriguingly, recent studies have

expression of certain claudins, especially claudin-1 andincreases during metastasis and genetic inhibition ofssion has a profound effect on the metastatic abilitieslls though in a tissue specic fashion [5254]. There is

lity that mutations in claudins may be causal to tumorHowever, to date there is no systematic sequence data

in any tumor type. On the other hand, gene silenc-promoter hypermethylation is a common feature ofers [55] and it has been suggested to underlie the down-f claudins in certain tumors. For example, a CpG islanded within the coding sequence of the claudin-4 gene,nt with a methyl-transferase inhibitor restored expres-protein in primary cultures prepared from high-gradedder tumors [56]. Furthermore, claudin-4 expressionted with its gene methylation prole in healthy anddders from 20 patients and claudin-6 expression is par-ed by promoter CpG island hypermethylation in MCF-7noma cells, while a synergistic effect of a demethylator

deacetylase inhibitors upregulates the expression ofs claudin-6, and sensitizes the cells for apoptosis [57].the mechanism by which decreased claudin expression

to the compromised TJ function and thus, neopla-to comprehend, but how increased claudin expression

to neoplastic progression is less clear. One plausible is that upregulation or aberrant tissue expression ofdins may contribute to neoplasia by directly altering TJd function. Furthermore, it is postulated that claudins

fect cell signaling pathways. Claudin proteins are likely signaling pathways via binding domains to ZO-1 at theirrminus [58].l adhesion proteins are known to play an important roleransformation when displaced from their normal mem-ization and could serve as oncogenic molecules, thed molecule being -catenin [59]. A similar functionality could be postulated for claudins, however, further



studies are needed to support such a notion. An increase in claudin-1 expression has been reported in human primary colon carcinoma,in metastasis samples and in the cell lines derived from primaryand metastatic tumors compared to their normal counterparts [54].Crucially, thcant subsetof liver mettion proteinoncogenic tTJ protein Zinduce dramSimilarly, gcancer cell ltural and fueffects uponathymic micatenin/Tcfunderlying complex in[54]. Expreulated by tsignaling pagenes regul

Metastaof specic ity and invClaudin-5 p(pro-MMP-of claudin-teinases (Talso promoMMPs and M(DeltaMT1-claudins msion and mobserved thincreased aclaudin-1 reSimilarly, ovincreased Mand protein10 overexpupregulatedcer cells mmembers [6

Most mainvade the potent targcuring infecegy for imprand basal mmation of cpolarity, wsis, leads toThe claudinmor therapthe high spworth mena precise tispecic bioing claudinall 158 ovacancers [67prognostic shown to bcolon cance

expression is differential between the subtypes and low versus highclaudin-1 expression helps identify highly aggressive triple nega-tive breast cancer [69]. Similarly, claudin-10 expression has beenshown to be an independent prognostic factor for hepatocellular

ma ntid/or

studd formbercale ch poain voplamorvia sdies gens.

upoeen ve bincluroacy sysely u

andproteme ibitinn of Pa proc to cerapsion ulating to

growion cy ma

llm

ractillulainterns co

- afnn in

thato forventson isng m

[9]. ood ahat allulaall pestrater, b

use metincluwnsxillin this article in press as: Jiang WG, et al. Tissue invasion and metastol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

ere was nuclear localization of claudin-1 in a signi- of colon cancer samples, particularly among the subsetastatic lesions. Nuclear localization of several cell junc-s (-catenin, ZO-1, ZO-2) is known to be correlated withransformation and cell proliferation [60]. Mutants of theO-1 that no longer localize at the plasma membraneatic EMT in Madin-Darby canine kidney I cells [61].

enetic manipulations of claudin-1 expression in colonines induced changes in cellular phenotype, with struc-nctional changes in markers of EMT, and had signicant

the growth of xenografted tumors and metastasis ince. Notably, regulation of E-cadherin expression and -

signaling emerged as one of the potential mechanismclaudin-1 dependent changes and thereby suggestedterplay between different cellcell adhesion moleculesssion of specic claudin family members can be reg-he wingless-type MMTV integration site family (Wnt)thway. Claudin-1 and claudin-2 are shown to be targetated by -catenin signaling [62,63].sis is a complex phenomenon that requires a numbersteps such as decreased adhesion, increased motil-asion, proteolysis, and resistance to apoptosis [64].romotes processing of pro-matrix metalloproteinase-22) by membrane type 1-MMP (MT1-MMP). Expression5 not only replaced tissue inhibitors of metallopro-IMP)-2 in pro-MMP-2 activation by MT1-MMP butted activation of pro-MMP-2 mediated by all MT-T1-MMP mutants lacking the transmembrane domain

MMP) [65]. It appears that interaction of MMP withight play an important role in tumorigenesis, inva-etastasis mediated by claudin expression. It has beenat overexpression of claudin-1 in colon cancer cellsctivity of both MMP-2 and MMP-9 while inhibition ofsulted in a signicant decrease in MMP-9 activity [54].erexpression of claudin-3 or 4 in ovarian epithelial cellsMP-2 activity [52]. An increase in mRNA transcription

expression of MT1-MMP was also observed in claudin-ressing cells, in which claudin-1, -2, and -4 were also, suggesting that the expression of claudin-10 in can-

ay dysregulate the expression of other claudin family6].lignant tumors are derived from, and most pathogensbody via the epithelium. The epithelium is therefore aet for improving drug absorption, treating cancer, andtious diseases. Modulation of TJ seals is a popular strat-oving drug absorption. TJs compartmentalize the apicalembrane domains of epithelial cells, leading to the for-ellular polarity. Loss of cellcell interaction and cellularhich often occurs in cancer cells during carcinogene-

exposure of TJ components on the cellular surface. family of proteins is an attractive target for antitu-y considering the epithelium-specic expression andecicity of claudin expression patterns in cancer. It istioning that claudin family members are expressed inssue-specic manner and thus could serve as tumormarkers. In this regard, a set of four markers, includ--3, was found to be sufcient to accurately identifyrian cancers tested, including eight early-stage serous]. Furthermore, claudin expression may be used as aindicator because high claudin-1 expression has beene associated with tumor progression and metastasis inr [68]. At the same time, in breast cancer, claudin-1

carcinothe idetify an(SAGE)alloweily melarge slish suto remand nethat tugeted antiboperfrinmarilyhave band hatypes tial appdeliveris widsurfacement, endosoby inhdomaiPSIF is specivant thexpresdysregsignalitumorinhibittherap

2.2. Ce

Inteextracewhich integri- andfering betweeeventsbut alslular eadhesiroundibarrierlikelihteins textraceies, smdemonHowevto be acanceraction and doand pa Molecular, biological and clinical perspectives. Semin

recurrence after curative hepatectomy [70]. Regardingcation of the claudin family of proteins as tools to iden-classify tumor types, serial analysis of gene expressionies of the breast [71] and ovarian [72] cancers have

the rst time the identication of specic claudin fam-s as potential biomarkers for these cancers. Althoughanalysis in a clinical setting will be required to estab-tential of claudins, basic research on claudins is likelyaluable for providing important insights into normal

stic cellular physiology. Preclinical studies have shown cells overexpressing claudins can be successfully tar-everal approaches, including the use of anti-claudinas well as the cytolytic enterotoxin from Clostridium

However, most of the studies have concentrated pri-n claudin-3 and claudin-4 [73]. Both of these proteinsidentied as targets of C. perfringens enterotoxin (CPE)een reported to be overexpressed in multiple tumording ovarian and prostate cancer. Yet another poten-h that has been suggested is the use of claudins as drugtem using Pseudomonas aeruginosa exotoxin A (PE). PEsed in cancer-targeting studies as it binds to the cell

is internalized via endocytosis. Following this, a PE frag-in synthesis inhibitory factor (PSIF), escapes from theto the cytosol [74], where it inhibits protein synthesisg elongation factor 2. PSIF lacks the receptor-bindingE, and fusion of a tumor antigen such as claudins withmising strategy for cancer-targeting therapy. Therapiesertain claudin family members could also serve as adju-ies. Highly increased and cytosolic/nuclear claudin-1in colon cancer has been reported [54,75] and claudin-1on modulates the balance between the Notch- and Wnt-

dysregulate colonocyte differentiation and promotesth and progression. Since Notch and or Wnt-signalingarries inherent high toxicity, the use of claudin-1 basedy provide an alternative.

atrix adhesion

on and adhesion between cells and the surroundingr matrix (ECM) classically involves cell surface integrinsact and bind ECM protein components [76]. Functionalnsist of a heterodimer structure made up of different

subunits and different integrin structures possess dif-ities for different matrix proteins [76]. The interactiontegrins and the ECM triggers a series of intercellular

not only results in the adhesion of the cell to the ECMms a mechanism for communication between intracel-

and the surrounding ECM. This process of cellmatrix essential for the attachment of cancer cells to the sur-atrix and subsequently the degradation of the matrixA number of integrins have been linked to metastaticnd cancer and/or stromal cells may deposit ECM pro-gain can enhance metastatic progression. Blocking ther part of the cellmatrix adhesion by means of antibod-ptides, and other natural- and phytochemicals has beened and has been covered by another article in this issue.locking the intracellular signaling event has also provedful approach in inhibiting this important event duringastasis. Key events following the matrixintegrin inter-de activation of the focal adhesion kinase (FAK), paxillintream chain signaling events [77]. Thus, inhibiting FAK

has become a hotly pursued approach in recent years.



CD44 represents another key cell adhesion molecule that holdspotential as an antimetastatic target both through its role in inter-acting with other cell types and the ECM. The CD44 gene, locatedat human chromosome 11p13, encodes the CD44s and CD44v iso-forms, whicisoforms shbinding sitas well as tfor ERM doand S6 kinanan receptomolecule [7of cancer srence, resistherapy [82target of canbody BIWAprodrug HYprodrug ONhyaluronancancer theadhesion msettling on and hence ntumors in t

2.3. Cellula

While eprocess of csis. Enhancis involved primary tumECM and inof cells requas the detecsurface procytoskeletobeen tightlyous proteinmigratory pnature [87,8treatmentsinvolved instrategy formal physiolwhere it is taken into cinhibit celluclinical sett

2.4. EMT

The proof morphomesenchymtransition. but has alsomotile cancination awaloss of cellthrough thteins such increase inand broneloss of E-ca

(cadherin switching) is a characteristic of EMT, seen in many can-cer types and is thought to account somewhat for the enhancedinvasive and motile properties of cancer cells [8]. Unsurprisingly,alterations in cell adhesion molecules (CAM) such as E-cadherin,

thed su

entiather tes cel me

an sion as desion atic

fromscripemoenotiatio (Tupreil, Sluated linkeolvesion mbeusseers o

otheamilrix-dy dyeen bseq

prootileinat

stratre likor ce

olecu

s nogregas, cocheson-morityasts,nd lymune thed thith or i

spon cellsts reptype moupothds onal, d this article in press as: Jiang WG, et al. Tissue invasion and metastol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

h arise through alternative splicing. CD44s and CD44vare the extracellular globular region that includes

es for hyaluronan, collagen, laminin and bronectinhe cytoplasmic tail region that includes binding sitesmain proteins (Ezrin, Radixin and Moesin), Ankyrinse related kinase (SRK). CD44 functions as a hyaluro-r, co-receptor for growth factors and as an adhesion882]. CD44 is involved in the malignant phenotypestem cells, including EMT, invasion, metastasis, recur-tance to chemotherapy and resistance to radiation85], which clearly indicates that CD44 is a potentialcer therapy. Humanized anti-CD44v6 monoclonal anti--4 (bivatuzumab), paclitaxel-conjugated hyaluronanTAD1-p20 (ONCOFID-P), SN-38-conjugated hyaluronanCOFID-S, hyaluronanirinotecan complex and other-conjugated drugs or siRNAs have been developed asrapeutics [86]. Therapeutics targeted to cellmatrixay represent a useful strategy to block cancer cells fromand subsequently penetrating vascular or cavity liningegatively impacting their ability to establish secondaryhe new site.

r migration

ssential to normal development and homeostasis, theellular migration is also a trait essential for metasta-

ed migration is key across the metastatic cascade andin the initial scattering of cells and migration from theor, the penetration of the basement membrane and

travasation and extravasation of vessels. The migrationires a number of intra- and extra-cellular events suchtion of extracellular signals by the cells, synthesis of cellteins and the coordination of intracellular signaling andn proteins. Throughout the literature, cell migration has

linked to cancer progression and metastasis. Numer-s and pathways have been implicated in altering theotentials of cancer cells and therefore their aggressive8]. Hence, given its essential role in cancer progression,

that inhibit cell migration or such proteins/pathways enhancing cellular motility represent an attractive

controlling metastatic dissemination. While in nor-ogy cellular migration is less active, there are processesessential, such as wound healing, and hence must beonsideration. Currently there are many compounds thatlar migration, although very few have been tested in aing.

cess through which epithelial cells undergo a serieslogical and biochemical changes to take on a moreal phenotype is known as epithelialmesenchymal

EMT is widespread throughout normal development been linked to the establishment of a more invasive,er cell phenotype facilitating detachment and dissem-y from the primary tumor [8992]. EMT involves thecell adhesion and the polarized epithelial morphologye characteristic loss of epithelial cell junctional pro-as E-cadherin, claudins and ZO-1, and a subsequent

mesenchymal markers such as N-cadherin, vimentinctin and cytoskeletal reorganization [91,93]. Indeed, thedherin and subsequent replacement with N-cadherin

impactsion anan esswith oregulacontrostratedexpreswork hexpresmetastcateninto tranbeen dmal ph

Initfactor ing in as Snaimplichence are invexpresin a nuas discmembact onMMP fin matquentlboth band su

Thesive, mdissempeuticcells aof tum

2.5. M

It iply agentitieapproaMany nat majbroblcular athe imacquirogy anthem wantitumcells recancertinue iphenoautonothe hydepenrection Molecular, biological and clinical perspectives. Semin

processes of cellcell adhesion and cellmatrix adhe-bsequently their metastatic potential. E-cadherin playsl role in the adhesion of cells and tissues and togethermembers of the adhesive complex, such as -catenin,ll adhesion, signaling and transcription in cancers and

tastatic progression [94]. Indeed, studies have demon-association between loss of E-cadherin and -cateninwith enhanced tumor cell invasiveness [95]. Othermonstrated an inverse correlation between E-cadherinand tumor cell invasion and motility and similarly withdisease in cancer patients [96]. The translocation of -

the adhesive structure to the nucleus, an event leadingtional activation of a number of target genes has alsonstrated to correlate with development of a mesenchy-ype [97,98].n signals, such as HGF, EGF and transforming growthGF-) are believed to onset the EMT process, result-gulation of EMT-inducing transcriptional factors suchg and Twist [99102]. Slug, Snail and Twist have beenin inuencing the expression of EMT proteins and ared to metastasis [103105]. For example Slug and Snaild in the down-regulation of E-cadherin [99,106] and thebetween Snail and E-cadherin is inversely correlatedr of cancers including breast cancer [107]. Similarly,d in Section 2.1.1, Snail exerts regulatory effects overf the TJ such as the claudins. These initiation factors alsor effector molecules to bring about EMT, such as they. Members of this family of proteinases play key rolesegradation, invasion, motility and adhesion and are fre-sregulated in cancer progression. Slug and Snail haveimplicated in the upregulation of MMP-2 and MMP-9uent EMT initiation [108].cess of EMT and subsequent acquisition of an inva-

phenotype with enhanced likelihood of invasion andion represents a key interest in cancer research. Thera-egies that can specically target this process in cancerely to be effective in reducing the metastatic potentiallls.

lar networks in the tumor microenvironment

w well established that solid tumors are not sim-tes of replicating neoplastic cells but are also livingmposed of numerous cell types, whose complexity, and may even exceed, that of normal healthy tissues.alignant cell types, referred to as the stroma, populate,

, the solid tumors. These non-malignant cells include resident epithelial cells, pericytes, myobroblasts, vas-mphovascular endothelial cells and inltrating cells ofe system. During malignant progression, neoplastic cells

ability to recruit, incorporate and reprogram the biol-e function of these healthy host cells, thus providingsupport, essential nutrients and weapons to hampermmune activity. In turn, the recruited non-malignantd by enhancing the neoplastic phenotype of the nearby, which again feed signaling back to the stroma to con-rogramming. Thus, the previous idea that the malignantof tumor cells was exclusively determined by cell-s genetic and epigenetic alterations is now replaced byesis that the malignant progression of cancer not only

tumor cells genetic aberrations but also on the bidi-ynamic and intricate network of interactions between

Please cite this article in press as: Jiang WG, et al. Tissue invasion and metastasis: Molecular, biological and clinical perspectives. SeminCancer Biol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008


Fig. 2. Cellular interactions within the tumor microenvironment. Numerous interactions between cell types are involved throughout tumor progression and metastasis.Communication between main components of the surrounding microenvironment play vital roles in enhancing metastatic potential, epithelial to mesenchymal transition(EMT), immune-evasion, mesenchymal to epithelial transition (MET) and angio- and lymphangio-genesis.



the cells of the stroma and cancer cells within the tumor microen-vironment [109,110] (Fig. 2).

Among the non-malignant cells that inhabit the tumormicroenvironment, cancer-associated broblasts (CAFs) and tumorinltrating-of the rolesgression. Cparacrine asive cancer a wide varECM-modiftion netwotogether goand metastprogressionporting angthe tumor-apressor celcells, regulasecrete chemDCs, TAMs athe expansimore tumothus leadininducing thtions, thus tumor micrvironment,progressionproliferatioremodeling

In cancepathways (transcriptiosecretion ofunction ofregulated o(RANTES)/Clates CAFs tumor-cell down-regulight polypeways, inltthus generaangiogenesimmune cerecruitmension. Tumorand bone mDCs to experate TH2 ctumor cell tumor cellsof immunoDCs surfacecells resultsproduction

Oncogencer cells triof several iIL-6, vasculchemoattrainduce thesecretion oenhance tumore it enh

and chemokines, including TGF- itself, IL-10 and CCL2/MCP1.TGF- and the potential for targeting this signaling pathway incancer is discussed in Section 4.2. A plethora of recent reportshas painted a consistent picture of how stromal cells (CAFs and

mato the e posic liferver, es, sP-9, Mthepting n thnd thichn by

colo furtting rived

liganhancmen. Thishem

like resuumoves r

medcturce ofsponce ofasis ls hasor ce

nodesed otacti

as th to inntians, ss nanviry CA

estaIL-10ntiatterises, Sattra

maucti

the e turnof IL-nvir

groalignmor n seris pu this article in press as: Jiang WG, et al. Tissue invasion and metastol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

immune inammatory cells are noteworthy because they play in tumor development and malignant pro-AFs secrete factors that act on tumor cells in bothnd autocrine fashions, thus resulting in a more aggres-phenotype. Across most cancers, activated CAFs secreteiety of growth factors, chemokines, collagens, andying enzymes, which collectively supply a communica-rk and an altered three-dimensional ECM scaffold thatvern proliferation of cancer cells and tumor invasionasis across tissue types. They also contribute to tumor

by recruiting tumor-promoting immune cells and sup-iogenesis. The tumor inltrating-immune cells includessociated macrophages (TAMs), myeloid-derived sup-

ls (MDSCs), dendritic cells (DCs), tumor inltrating Ttory T cells (Tregs) and mast cells [109]. Tumor cellsokines and cytokines that are able to recruit mast cells,

nd MDSCs. Tumor cells also activate mast cells, promoteon of the MDSCs and the polarization of TAMs. Further-rs both inhibit DC maturation through IL-10 secretion,g to antigen-specic anergy, and reprogram the DCs,em to exert immunosuppressive or angiogenic func-resulting in an immunosuppressive and inammatoryoenvironment. Once recruited to the tumor microen-

these immune cells can contribute to the malignant of the cancer-cell phenotype by supporting tumorn, survival, invasion, metastasis, angiogenesis and ECM.r cells, the constitutive activation of various signalingincluding MAPK, signal transducer and activator ofn 3 (STAT3) and -catenin pathways) results in thef cytokines which modulate the recruitment and

the stromal cells. In particular, the tumor-derivedn activation, normal T cell expressed and secretedhemokine (CC motif) ligand 5 (CCL5) cytokine stimu-to externalize the S100A4 protein, which stimulatessurvival and migration, up-regulation of the MMPs,lation of TIMPs, activation of the nuclear factor of kappaptide gene enhancer in B cells (NF-B) and MAPK path-ration of T cells and nally, up-regulation of RANTES,ting a signal amplication loop. RANTES also induceis and act as chemoattractants for additional effectorlls. Tumor-derived stem cell factor (SCF) promotes thet and activation of mast cells and the MDSC expan-s also secrete the thymic stromal lymphopoietin (TSLP)arrow stromal cell antigen 2 (BST2). TSLP induces

ress OX40 ligand, which directs CD4+ T cells to gen-ells secreting IL-4 and IL-13. These cytokines preventapoptosis and indirectly promote the proliferation of

by stimulating TAMs to secrete EGF. BST2 is a ligandglobulin-like transcript 7 (ILT7), which is expressed on. The interaction of ILT7 on DCs with BST2 on tumor

in inhibition of IFN- and pro-inammatory cytokines by DCs with immunosuppressive effects.e activation and subsequent signal activation in can-gger multiple cascades thus resulting in the secretionmmunosuppressive molecules, including TGF-, IL-10,ar endothelial growth factor (VEGF), CCL2/monocytectant protein 1 (MCP1), cyclooxygenase-2 (COX2), that

immunosuppressive immune cells. Production andf these factors by both cancer and surrounding cellsmor cell proliferation, migration and invasion. Further-ances the production of immunosuppressive cytokines

inamwithinprovidand bacell proMoreoenzym7, MMand capromoinvasiocells aEGF, winvasiosion ofwhichgeneracell-demotif)can enrecruittumorupon cis mostwhichin the talso gialso bearchitepresena correnicanmetastcer celof tumlymphexpreschemoas wellabilitypreferefunctiorecruitmicroealso bfor themore, differecharaccytokinchemoare thethe indble forTGF-levels microe

Thenon-ming tucells caFor th Molecular, biological and clinical perspectives. Semin

ry cells) can promote malignant progression. Indeed,primary tumor microenvironment, the stromal cells

tent oncogenic signals, such as TGF-, HGF, EGF, Wnt,broblast growth factor (bFGF), which stimulate cancer-ation, survival and invasion, thus facilitating metastasis.these cells produce several angiogenesis-modulatinguch as VEGF, thymidine phosphorylase, MMP-2, MMP-

MP-12, COX2, urokinase plasminogen activator (uPA)sins B and D, which together degrade the ECM, againmetastasis. TAMs promote carcinoma-cell motility andrough a paracrine signaling loop between the tumorhe TAMs. Within this loop the macrophages express

promotes formation of elongated protrusions and cell carcinoma cells. In addition, EGF promotes the expres-ny stimulating factor 1 (CSF-1) by the carcinoma cells,her promote the expression of EGF by macrophagesa positive-feedback loop. The secretion of stromal-

factor 1 (SDF1), also known as chemokine (CXCd 12 (CXCL12), by TAMs and CAFs at a tumor site

e the invasion, intravasation, metastasis formation andt of MDSCs, TAMs and endothelial cells to the primary

enhancement of invasion and intravasation dependsokine (CXC motif) receptor 4 (CXCR4) signaling, and itly to occur through activation of CXCR4 on macrophages,lts in increased paracrine interactions with tumor cellsr microenvironment. Increased CXCL12/SDF1 secretionise to an increased microvessel density, which mightiated by TAMs and might contribute to an altered tumore, thus resulting in increased intravasation through the

a higher density of entrance sites into the blood, withding increase in the formation of metastases. The sig-

CXCL12/CXCR4 signaling in breast cancer invasion andis widely appreciated. CXCR4 expression in breast can-

been shown to increase metastasis through the hominglls to sites of increased CXCL12 expression, such as thes. Similarly, the interaction of CXCL12/SDF1 and CXCR4n mammary adenocarcinoma MTLn3 cells increases thec and invasive behavior of these cells to CXCL12/SDF1,eir motile behavior within the primary tumor and theirtravasate. TAM-derived CCL17 and CCL22 chemokineslly attract T cell subsets that are devoid of cytotoxicuch as Tregs and Th2 lymphocytes. TAM-derived CCL18ve T cells, which induce T cell anergy. Within the tumoronment IL-10, secreted not only by immune cells, butFs and tumor cells, is the main cytokine responsibleblishment of the immunosuppressive milieu. Further-, together with IL-4, IL-6 and IL-13, induces monocyteion toward a mature M2-polarized phenotype that istic of TAMs. At the tumor site, the IL-1 and IL-6100A8 and S100A9 pro-inammatory proteins and thectant molecules CCL2/MCP1, CXCL12/SDF1 and CXCL5in factors that are responsible for the recruitment andon of MDSCs. VEGF is one of the main factors responsi-xpansion of MDSCs, while IL-4, IL-13, IFN-, IL-1 and

on their suppressive functions. MDSCs produce high17, which further exacerbates the inammatory tumoronment.wing body of evidence regarding the roles played byant cells of the tumor microenvironment in promot-progression indicate that it is conceivable that theseve as novel therapeutic targets in the cancer treatment.rpose, several therapeutic approaches that use small



Table 1Effects of approved and experimental targeted agents on tumor cells and tumor microenvironment stromal cells.

Drug Drug class Target Effect on tumor Effect on the immune system References

STI571 (Gleevec orimatinib m

Small molecule PDGFR and Reduces microvessel density Prevents mast cell proliferation [283,284]

Bevacizuma

IM-2C6 SU5416

MMI-166 MMP-genes

S-3304 P-9

Dasatinib kemic

Dipyridamo and

Bindarit and

Upanap-126

ATN-658

L2G7/Rilotum

Trastuzumab

Cetuximab

MGA271

AMD3100

Celecoxib

5-Fluorourac

All-trans-ret(ATRA)

Sclareol Temozolomi

molecule intarget molement, activcells have bsummarizeuse in cancagainst tumeral strategfactors (e.gclinical stuefcacy in tSimilarly, seoped to taanti-angiogcells and inSTI571 andesylate) inhibitor c-Kitb Monoclonal

antibodyVEGF Blocks angiogenesis

Antibody VEGFR Blocks angiogenesis Small moleculeinhibitor

VEGFR Reduces vascular density

Small moleculeinhibitor

MMP-2 andMMP-9

Suppresses MMP-2 and activities; inhibits angioand tumor growth


MMP-2 andMMP-9

Inhibits MMP-2 and MM


c-Kit, ABL, SRC,PDGFR

Induces apoptosis in leu

le Small molecule Wnt, MAPKand NF-Bpathways

Decreases tumor growthmetastasis

Small molecule CCL2/MCP1 Decreases tumor growth this article in press as: Jiang WG, et al. Tissue invasion and metastasis:ol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

metastasis RNA aptamer uPA Delays the proteolitic convers

of pro-uPA to active uPA; inhibtumor cell invasion; reduces thtumor cell intravasation anddissemination

Monoclonalantibody

uPA receptor(uPAR)

Decreases tumor cell invasionmigration and tumor volume

umab Monoclonalantibody

HGF Inhibits the tumor growth

Monoclonalantibody

HER2 Blocks growth signals

Monoclonalantibody

EGF receptor(EGFR)

Blocks growth signals

Monoclonalantibody

B7-H3 NA

Small molecule CXCR4/CXCL12(SDF1)signaling

Sensitizes cancer cells tochemotherapy: inhibits tumorgrowth


COX2 NA

il Small molecule Thymidylatesynthase

Promotes the cytotoxicity of tucells

inoic acid Vitamin Aderivative

NA NA

Phytochemical NA Decreases the tumor size de Small molecule DNA Promotes the cytotoxicity of tu

cells

hibitors, antibodies or phytochemicals that specicallycules and signaling pathways involved in the recruit-ation and function of tumor inltrating non-malignanteen tested in both animal models and human. Table 1s the most up-to-date drugs available with potentialer therapy, known effects on tumor cells and activityor-stromal microenvironment communications. Sev-ies to inhibit either CAF activation or CAF-derived. HGF, uPA, CXCL12/SDF1) have been applied in pre-dies of cancer therapies and the results have shownhe inhibition of tumor growth and invasion [111116].veral immunotherapeutic approaches have been devel-rget immune cells that inltrate the tumor. Someenic agents impair proliferation and survival of mastduce DC maturation and their antitumor activity (e.g.

bevacizumab). The impairment of the stem cell factor

(SCF)/c-Kit both tumorstrategies thsuppressivehuman. Theinduction oretinoic aciinltration their immuuorouraci[119] decrebreast and Finally sclanumber of further studbest therapand survivalIncreases DC maturation, shifts DCdifferentiation toward mature DCsinstead of MDSCs and increases DCpriming of T cells

[285,286]

NA [287]NA [288]

9is

NA [289]

NA [290]

cell Induces apoptosis in mast cell [117]

Decreases TAM and MDSCinltration

[118]

Decreases TAM and MDSC [119] Molecular, biological and clinical perspectives. Semin

inltrationionitse

NA [111]

and NA [112]

NA [113,291,292]

Primes antitumor CTLs, boosts NKsecretion of IFN- and mediatespotent antibody-dependentcellular cytotoxicity

[293]

Immune activating: increases MHCclass I and MHC class II expression;augments DC priming oftumor-specic CTLs.

[294]

Mediates potentantibody-dependent cellularcytotoxicity

[295]

Reduces the recruitment ofbone-marrow derived cells

[114116]

Decreases both MDSC numbersand function

[296]

mor Induces MDSC apoptotic cell death [297]

Reduces MDSCs [298]

Decreases the number of Tregs [120]mor Reduces the number of Tregs [121]

signaling pathway by dasatinib induces apoptosis of cells and mast cells [117]. Several immuno-therapeuticat target MDSCs and that can neutralize their immuno-

effects have been reported in both animal models andse strategies include approaches that are aimed at thef differentiation of these immature cells [e.g. all-trans-d (ATRA)], or of a decrease in their number and tumor(e.g. dipyridamole and bindarit), or at interfering withnosuppressive functions (celocoxib), or killing MDSCs (5l- or 5FU). Interestingly, dipyridamole [118] and bindaritase the inltration not only of MDSCs but also of TAMs inprostate cancer proof of concept animal model studies.reol and temozolomide reduce tumor growth and thetumor inltrating Tregs [120,121]. Therefore, althoughies will be needed to determine which cell(s) is/are theeutic target(s) and which drugs are the most efcient



and selective, there is no doubt that the therapeutic targeting oftumor microenvironment cells represents a valuable strategy tocomplement conventional anticancer strategies.

2.6. Cancer

Cancer stroversial carcinogeneThe CSC repcells that caThese highthe growthto be resistresponsiblerst isolatedinvestigatiotumorigeniincluding b

The critito be: tumerogeneity subpopulatthe other cpopulation of cells surfisolate the CCD44, CD26dehydrogencombinatioevery tumohas been foprognosis aCSC theorythat can prthe develop

In breasand low levshown to eAnother malevels ALDHpopulationsALDH+ andnature wasthe gene endescribed tmetastasis,cer cell lineALDH cellcombined isible role foCSC. Additiextra-pulmney [129]. Swhere the to be essenin an orthosupports thmetastasis.

There iwithin theenhanced pfor metastathe pancreaexpression.lation of m

CXCR4 is a protein that has previously been implicated in can-cer cell metastatic potential [132]. Using an orthotopic model ofpancreatic cancer only the migrating CSC population, expressingCD133+CXCR4+, were able to establish liver metastasis. Inhibition

R4 siresulf thee of Citionsibleeporting l), aning Lute ed f they tum

tum ands, hoasis ore spmeT has

cells to e

canetaso EM

to Cntly ing

tissuls unCSC end Sl

CSC prefs andeen dnin, e

the nt ha

cells fromnt exly retz et T hachymal c

ay rte b

ed totastpmehanco theat re

typeent ludincer chymduc this article in press as: Jiang WG, et al. Tissue invasion and metastol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

stem cells (CSC)

tem cells (CSC) present an exciting yet somewhat con-eld in cancer research. In the cancer stem cell model ofsis there is a hierarchical organization of cancer cells.resent a highly tumorigenic sub-population of cancern be isolated from other cancer cells in the same tumor.ly tumorigenic cells have been proposed as crucial to

and development of primary tumors and are believedant to conventional therapy and therefore likely to be

for disease recurrence and treatment failures. CSC were in acute myelogenous leukemia (AML) and subsequentns of solid tumors have revealed the presence of highlyc cancer cells (CSC) in essentially every solid cancer typereast, lung, colon, pancreas, head and neck [122127].cal characteristics of a CSC require these cancer cellsorigenic, able to reproduce the original tumor het-including both the tumorigenic and non-tumorigenicions of cancer cells, self-renewing, and separable fromancer cells. CSC typically represent only a small sub-(



Although attractive and in agreement with many genetic anal-yses and with the evaluation of CSC in primary tumors and animalmodels, to date the causative role of CSC in metastasis formation hasnot been formally proven. Metastases represent one of the key fac-tors in treatCSC play a ctant step todevelop. Thhave not yetigation. Thand improv

3. Cancer c

3.1. Organ

The premetastatic ent for centand soil thecic tumor organ or locpropensity example, brmetastasizicancers freqa few organcancers, namas the spleeing this orgby many castudied wittors may ag

3.1.1. TargeBone me

of cancers, rence of ththe qualitymorbidity. cant bone dbone formainvolved incells in theevents thatprocesses care more sbone enviroosteomimicThe currentstanding, trmetastatic d

The spremarrow invstromal celis played bmotif) ligantion to the correspondand its receof cancer c3100, T140decreasing els of breasproteins (b

can stimulate bone matrix invasion by binding the surface of tumorcells through specic membrane receptors such as integrins V3and 21 [145] and it has been demonstrated that breast cancercells expressing V3 integrin and prostate cancer expressing

2, hlinict bo

cellscer eatic n anasis as btes asis sed issionXL18

signowthmize

resiion-rized

reate22 pse ints had impl (PFo (p 85%) for accurately diagnosingional lymph nodes, but only those which are com-aced by tumor cells (i.e. metastatic lymph nodes, nottic disease). Precision in the preoperative detection of

metastases is of great importance as the trend towarded cancer care and minimally invasive surgery gath-tum. Second, the not infrequent observation of laterrrence in patients who have seemingly had a com-ion of their tumor suggests that clinically undetectablesits must be present at the time of operation and themph nodes are a frequent site of tumor recurrenceat this compartment must be an important site for

ase. Recent studies indicate that 117% of histologi-ive lymph nodes, and 1150% of pathologically nodetients have nodal metastases that were missed by rou-gic examination [176,177]. This therefore means thatN designation is often incorrect and results in subopti-ent decisions. Robust, sensitive immunohistochemicalusing antibodies to detect epithelial tumor cells in

tissue have been in use since the mid-1990s [178].re surprising that no consensus exists regarding thesignicance of immunohistochemically identied iso-

cells (ITCs) in many tumor types [179194]. The mainthis is the lack of unequivocal results showing theirsignicance in various solid tumors. However, manyfer from small numbers of patients, limited analysisparafn blocks, and, most importantly, varying de-both isolated tumor cells and micrometastases. As pertion of the AJCC cancer staging manual, micrometas-ccult metastases greater than 0.2 mm but not greater

in size, while ITCs are dened as small clusters of cellsthan 0.2 mm, or nonconuent or nearly conuent clus-

not exceeding 200 cells in a single histologic lymphsection [175]. In esophageal cancer, it was found thation between an isolated tumor cell and a micrometas-ot important [195]. Patients with either of these ine lymph node(s) had signicantly reduced overall sur-ared to patients who remained node negative afterns and immunohistochemistry. The importance of iso-r cells in lymph nodes has been reported not only in

cancer [181,184187,189], but also in several stud-ic cancer [196,197], melanoma [198], breast cancercolorectal cancer [203205], and non-small-cell lung]. Lymph nodes containing isolated tumor cells shouldgnated pN0(i+) as per the AJCCs breast cancer stag-. Whether these cells represent tumor cells in transit, but they are associated with a worse prognosis com-re likely true node negative (pN0) patients. It is likely

cells represent microscopic tumor cell dissemination,l and economic constraints often prevent their routine

we often categorize cancer as either localized orthis simplistic thinking might be misleading. Accord-

et al. [207], the true nature of the disease mightnceptualized within an evolutionary model, in whichous selection of genetically unstable variant cells andsion determines disease course and risk of dying from

therefoprimarsettingMortona primin cancremainregardHowevconcepeconomhistopwill prers, whdeliverwithinin receprogrewith hagentsanti-CDoptimaand inand trsignicone stphotrosystemtumorin the this aplymphwith eamay ofease, isystemcarbonorescenodes nique this apfuture staging

If cdeposition wtissuesorgan tems fpolymmost ptems adesiredthetic them vtonealevidenall dienologyand funano-/treatm Molecular, biological and clinical perspectives. Semin

splay quite different chromosomal aberrations from theor, and require specic targeting in adjuvant therapy

e sentinel lymph node (SLN) concept, rst described byhe early 1990s [208], depicts the preferential drainage ofmor to a regional lymph node(s). It is the gold standardre for patients with breast cancer and melanoma, buttroversial in other solid tumors with continued debates role, if any, in staging and treatment algorithms [209].ere are 2 reasons why all cancers should adopt the SLNst, SLN biopsy is the only practical method in todayslimate to identify the most important nodes for detailedogical analysis. And second, adoption of this techniquete the development of novel sentinel lymph node trac-are capable of non-invasive lymph node staging, andchemotherapeutic agents to disseminated tumor cellsnodes. Many novel nanomaterials have been proposedars for medical applications [210] and they are rapidly

toward clinical medicine. Nanoparticles (1030 nm)binding afnity for lymphoid cells are ideal imaginghe SLNs. Working toward this hypothesis, conjugatedntibodies with gold nanoparticles (18 nm), through anyethylene glycol (PEG) coating, have been constructedd into mice [211]. Analysis conrmed rapid uptakeort of the nanoparticles in the lymphatics, as well asretention in the lymph nodes. Taking this applicationrther, Weissleder et al. [212] have shown that lym-

superparamagnetic iron oxide nanoparticles, injectedy as exogenous contrast, can discriminate healthy versusened nodes by the degree of accumulation of particless. At present, there are limitations to the accuracy ofch, as false negative results may occur in the case ofs less than 5 mm in diameter, and by extension, thoseicrometastatic disease. An alternative approach, which

etter accuracy when dealing with micrometastatic dis-inject the tracer via the interstitial route rather thany. Recent studies have shown that mammaglobin-A orhydrase specic mAbs conjugated to near infrared u-es can detect as few as 1000 cancer cells in the lymph

interstitial injection [213,214]. Although this tech-rrently limited to animal studies, the application ofch to other imaging modalities holds promise for thelopment of reliable, accurate non-invasive lymph node

divergence does indeed exist between occult tumor lymph nodes and the primary tumor, a logical solu-

be to deliver anticancer drugs directly to lymphaticich would optimize response and limit nonspecicities of systemic chemotherapeutic agents. Carrier sys-argeted lymphatic delivery include liposome-based,sed, and immunotherapy-based [215]. Perhaps theising of these are the polymer-based drug delivery sys-ticle size can be carefully standardized to achieve thect. As an example, Liggins et al. [216] loaded a syn-mer microsphere with paclitaxel (PTX), and injected

intraperitoneal route into a rat model with intraperi-r cells. Rats treated with PTX microspheres showed no

tumors in the peritoneal cavity, while those without,ithin 4 weeks. With recent advances in nanotech-

a better understanding of the lymphatic anatomyn, targeted chemotherapeutic delivery via polymericoparticles may greatly improve efcacy of anticancer.



3.2.3. Transcoelomic metastasisSpreading and formation of metastatic tumors in the body

cavities (mostly in peritoneal and pleural cavities) is a commonfeature in cas transcoefrom tumormetastasis and ovarian[217]. The mthe peritonmain cell tyMalignant tlioma, an alittle effectimay be a pthought to by mesothea protectivoccurs via oimplantatiocavities aretemic routeand lung catumors orignamely tumbladder. Casues, breacin the peritsub-serosalperitoneal stial step in

When tulesion in thto adapt thmesotheliadevelopingsis does nota wider sprcers whichperitoneal tric cancer [220,221] aperitoneal moneal metasoptions for involves suever, this hIn the caseintraperitonprimary caManagemeand early incal debulkithis procedeal cavity, have tried tseeding arepresents ansis and to acarcinomattion. Apart surgery (foavoid contapracticed asurgical prosibly existi

and further research is essential in order to develop alterna-tives.

The contact of cancer cells to mesothelial cells is followed con-tly by adhesion, invasion and growth of tumor cells at suchsite. d grtizehesionanolectors aition

cell olec

yalurfactoovid

cellsance

to gationisms, an

rape

atura

date,ientsceduer, caat hant sctionnal iity [2mendreen(radiade etasch tial p

as wered fns. to t

c cele monew natioghligs [22y 5%operre tably, e aducingnal deded. For

hea this article in press as: Jiang WG, et al. Tissue invasion and metastol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

ertain malignant conditions and are broadly referred tolomic metastasis. Trancoelomic spread occurs mostlys to adjacent tissues/organs. Transcoelomic peritonealarise mostly from pancreatic cancer, colorectal cancer

cancer, followed by gastric cancer and cervical canceresothelium is the lining of cavities in the body, mainly

eal cavity, pleural cavity, and pericardiac cavity. Thepe that forms the mesothelium is the mesothelial cells.ransformation of mesothelial cells results in mesothe-ggressive malignant condition against which there isve treatment. It has been reported that mesothelial cellsrivileged site for tumor cells to attach [218]. This wasbe due to the layer of hyaluronan, a molecule releasedlial cells and which, together with other proteins, formse surface on the mesothelium. Peritoneal metastasisne of two main routes: systemic spreading and localn after invasion of local tissues. Tumors away from the

likely to develop transcoelomic metastasis via the sys-, for example, peritoneal metastasis from breast cancerncer. Perhaps most peritoneal metastases come frominated from organs adjacent to the peritoneal cavity,ors from the stomach, colon, pancreas, ovaries, and

ncer cells from these tumors invade surrounding tis-h the peritoneal lining and spread by way of seedingoneal cavity, although trans-serosal, inter-serosal and

spread can also be seen. It is clear that in most forms ofpreading, tumor-mesothelial interactions are an essen-establishing a metastatic tumor in the cavity.mor cells disseminate through and develop a metastatice pleural or peritoneal cavity, the cancer cells needemselves to the environment and interact with thel cells. Certain tumors have a far higher incidence of

peritoneal metastasis. Of course, peritoneal metasta- occur alone, and can be seen as locoregional issues ofead of cancer cells. For example, 70% of ovarian can-

have local regional lymphatic metastases also havemetastases [219]. Similarly, 50% of patients with gas-which has invaded serosa have peritoneal metastasisnd the majority of patients with pancreatic cancer haveetastasis [222]. Sadly, patients with wide spread perit-tasis survive no longer than 6 months [223]. Treatmentperitoneal metastasis are rather limited. Managementrgical procedure to remove the primary tumor. How-as little impact on established peritoneal metastasis.

of systemic metastasis, systemic chemotherapy andeal chemotherapy have been attempted to treat the

ncer and peritoneal metastasis with limited benet.nt of peritoneal metastasis also involves preventiontervention. A critical opportunity is during the surgi-

ng of the primary tumor, a key procedure. However,ure may also introduce tumor cells into the periton-although this has been a consequence that surgeonso avoid. In addition, peritoneal metastases or tumor cell

likely to exist at the time of operation. Surgery itself excellent opportunity to prevent peritoneal metasta-ct early on the metastasis before it becomes full scaleosis. Yet, there are very few options for this interven-from techniques to prevent articial seeding, duringr example the use of padding/isolation materials toct between tumor and surrounding tissues), a widelypproach is peritoneal washing/irrigation following thecedure with the aim to remove any debris and pos-

ng tumor cells. This is hardly a satisfactory solution,

sequena new sion antraumacell adhyalursion mmediain condtumorCD44 mwith hhand, also prcancerfrom cstromainvestimechacavitie

4. The

4.1. N

To for patcal prohowevcells thimportthe funfunctiomortal

Trecer scchemobeen mfrom mapproaessentgrowthdrugs advancregimeadded speciincludthese combiThis hitumor

Onlinal prdrugs aInvariathey aron redtigatiomust rburdenwound Molecular, biological and clinical perspectives. Semin

The exact role of mesothelial cells in tumor cell adhe-owth is unclear. Many studies have demonstrated thatd mesothelial surfaces are privileged sites for tumoron possibly due to the binding of tumor cells to the

coat of mesothelial cells [218], upregulation of adhe-ules on mesothelial cells in response to inammatorynd exposure of underlying ECM. However, hyaluronaned medium from cultured mesothelial cells prevented

attachment to mesothelial cells, possibly by binding toules on the tumor cells and preventing their interactiononan on the mesothelial cell surface [224]. On the otherrs released from tumor cells or adjacent stroma maye a favorable environment for the interaction between

and mesothelial cells. For example, IL-1 or TGF-1r cells can act on the mesothelial cells and/or adjacentpromote peritoneal dissemination [225,226]. Furthern into this particular interaction will shed light on the(s) of cancer cell dissemination in pleural and peritoneald may also provide novel therapeutic opportunities.

utic approach to cancer metastasis

l products with antimetastatic properties

surgery remains the primary cancer treatment option who are deemed curative at diagnosis. Existing surgi-res are successful in removing the majority of tumors,ncer cells that were missed during surgical removal or

ad already migrated out of the primary tumor sites areources for metastasis. The migrated cells later impair

at the newly metastasized organ sites. Eventually, thempairment at metastatic sites results in cancer related27].ous advancements have been made in can-ing, early diagnosis and development of novelo)therapy regimens. However, little progress hasin cancer prevention or containment of primary tumorstasizing. Therefore, there is a need for a multiprongedo prevent the primary tumor from spreading. Tworoperties of new anticancer drugs are to stop tumorwell as inhibit metastasis. Historically, chemotherapye developed to manage primary tumors. We haverom using single drugs to combination chemotherapyIn the past decade, several new agents have beenhe chemotherapy arsenal. These new drugs targetl signaling pathways. Some of the new targeted agentsnoclonal antibodies and kinase inhibitors. However,agents are not effective by themselves, but only inn with other antimetabolite drugs like 5-uorouracil.hts the need for more drugs that could target primary8].

of small molecules (investigational drugs) with medic-ties enter human clinical trials. Several investigationalken off clinical studies due to toxicity or lack of efcacy.the trial drugs are seldom used as single agents. Instead,ded on to existing drugs. Moreover, clinical trials focus

primary tumor burden. On the other hand, these inves-rugs may potentially inhibit metastasis. Therefore, weicate our efforts to go beyond reducing primary tumor

example, EMT, while important in development andling, is detrimental in cancer patients as it is a hallmark



of metastasizing cancer cells. Thus, focus on how to prevent EMTin cancer cells represents a key area of research toward treatmentdevelopment. A good approach could be to develop a set of mark-ers affected in EMT. Such a panel could serve as a benchmark forsmall moleerties. Usuamay have akey EMT regscreens maycancer patie

Many annatural prosources forapy agents not all the bintroduced or do not insmall moletumor promon tumor ceactivities oatically screas well. Mocytotoxic dr

Drug dislenges. Theis biodiversmolecules icle. Even whsynthesis oably, enviroaffect acquViewed frorules and ridenticatioget cancer mongoing andactivity.

This reviinvolved intype and mprocess fullthermore, idinterfere werating a nIdeal compspecic effeeffects acrosubsets andthis reviewsent key aretherapeuticGiven the cthat of tissand potentiinvestigatedprogressionto highlightgets (Table approachesacross the licise for crouseful inforcomplemenareas and, ttumor-stim

4.1.1. SilibininOne such example is seen in the antimetastic properties of

silibinin, a plant derived anticancer agent. Silibinin is a mixtureof two avonoids, silibinin A and silibinin B. It is derived from

lk thatopstrat30]. /Apooweigna

indstrate SK-show

levecle 4B-2

2, phnalin

Yangditiontiae keave hengncer

[235, didoug

been canctal c

cells lun

verier cibitouentrgetsueny oner stu

canatrix

dep durar enuent

FAKatrixibitotreatof DM

the nd lu

perior grritonaininnhibn ofial atast this article in press as: Jiang WG, et al. Tissue invasion and metastol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008

cular screens to select molecules with anti-EMT prop-lly, these small molecules alone are not cytotoxic butntimetastatic activity through their interactions withulators, etc. The small molecules selected through such

be combined with the standard of care drugs to benetnts.ticancer agents in use were originally developed fromducts. Plants, fungi and marine organisms are the major

new drug discoveries. About 6085% of chemother-in use are natural product derivatives [229]. However,ioactive molecules isolated from a natural product areto the clinic because these molecules may be either toxichibit cancer cell growth in vitro. However, the isolatedcules may inhibit specic signaling factors involved inotion. Current anticancer drug discovery efforts focusll toxicity. Such approaches will miss specic inhibitoryf the test compounds. Therefore, we need to system-en small molecules for their antimetastatic propertieslecules thus identied, may be combined with otherugs to inhibit metastasis.covery from natural products has two important chal-

rst is technical. The second, and equally important,ity (governmental) regulations. Synergism of multiplen crude preparations is an important technical obsta-en a single active molecule has been identied, in vitrof natural products has proved difcult. Understand-nmental concerns and intellectual property (IP) issuesiring natural products nationally and internationally.m a broader perspective, the Biodiversity conventionegulations are essential. Despite these obstacles, then and generation of new therapeutic strategies to tar-etastasis has great potential. Research into this area is

numerous compounds have demonstrated anticancer

ew outlines and has discussed key factors and pathways the establishment of an invasive cancer cell pheno-etastatic dissemination. The ability to understand thisy is an invaluable tool in combating this process. Fur-entication of suitable, low-toxicity compounds which

ith these processes in cancer cells is paramount in gen-ew generation of cheap, readily available compounds.ounds would have low inherent toxicity with cancercts, would have low cost and be readily available, havess a broad range of cancer types rather than specic

be free of intellectual properties. In respect to this, outlines a number of possible targets that may repre-as to target metastatic spread combined with potential

approaches to interfere with such targets (Table 2).omplex nature of cancer which extends beyond justue invasion and metastasis, these target approachesal strategies have also been explored and their efcacy

in the key hallmark areas of cancer development and (outlined in the other articles of this journal edition)

potential overlaps and further illustrate how these tar-2) and approaches (Table 3) may have benecial holistic

to cancer. Together these tables summarize key dataterature. This cross-validation is a very important exer-ss-target and cross dicipline verication. It providesmation as to whether these approaches and targets havetary or contrary interactions with the other hallmarkhus, their likelihood of resulting in pro-carcinogenic orulating effects.

the miits hepdemontasis [2(TRAILway. Hdeath scells bydemoncell lininin is mRNAsion cyMDA-MVEGFRthe sig

4.1.2. Tra

of potesion arTCMs hYangzting cain vivoDME25cells thIt has gastriccoloreccancertion ofit wasof cancan inhconseqalso taThe inon onlAnothetion ofcell matrationcriticalvasculsubseq

Thecellman inhwhile effect stratedcells amouseof tumintrapecent stcould ibinatiopotentthe me Molecular, biological and clinical perspectives. Semin

istle plant Silybum marianum. This plant is known forrotective properties. In vitro and in vivo, silibinin ised to inhibit cancer cell migration, invasion and metas-Tumor necrosis factor related apoptosis inducing ligand2L) is an important mediator of intrinsic apoptotic path-ver, some tumors fail to respond to TRAIL mediated cellls. Silibinin induces apoptosis in TRAIL-resistant tumorucing the caspase cascade [231,232]. Youse et al. [233]ed that silibinin inhibited the growth of neuroblastomaN-MC by downregulating Akt-mediated NF-B1. Silib-n to inhibit metastasis by inhibiting the expression of

ls of GDP dissociation inhibitor (D4-GDI) and cell divi-2 (Cdc42) in the highly metastatic breast cancer cell line,31 and protein levels of CD31, nestin, VEGF, VEGFR1,ospho-Akt and hypoxia-inducible factor-1 (HIF-1),g molecules involved in neovascularization [234].

zheng Xiaojinal chinese medicines (TCMs) represent another sourcel antimetastatic agents. Cancer cell adhesion and inva-y traits in the metastatic cascade. While relatively fewbeen reported to inuence these steps, the formulation

Xiaoji (YZX) has demonstrated an efcacy in inhibi- cell adhesion, migration and angiogenesis in vitro and237]. YZX capsules consist of 16 herbs. An YZX extract,

not show a signicant effect on the growth of cancerh it markedly suppressed cell adhesion and migration.

demonstrated that YZX inhibited the cell adhesion ofer cells (HGC27) in a concentration dependent manner,ancer cells (HRT18), breast cancer cells (MCF7), lung

(A549) and osteosarcoma cells (MG-63) and the migra-g cancer cells and colorectal cancer cells. In addition,ed that the inhibitory effect of YZX on the adhesion

ells is related with PI3K signal pathway. Wortmannin,r of PI3K activity, can suppress PI3K/AKT signaling andly reduce adhesiveness of cancer cells. DME25 which

the AKT pathway can enhance such inhibitory effect.ce of DME25 on the PI3K pathway may not depend

signal pathway, namely the AKT pathway [235,237].dy has demonstrated that YZX can suppress the forma-liculus of vascular endothelial cells, and indicated thatadhesion and migration could be inhibited in a concen-endent manner [236]. Cell adhesion and migration areing angiogenesis, particularly canaliculus formation bydothelial cells when they adhere to the cell matrix andly migrate in the ECM.

signaling pathway is a key pathway involved in adhesion [238240]. DME25 has been reported to havery effect on the phosphorylation of the FAK pathway,ment with a FAK inhibitor signicantly enhanced theE25 on the FAK pathway [241]. YZX has also demon-

ability to not only inhibit the growth of colorectal cancerng cancer cells but also to suppress the formation oftoneal tumor nodules in vivo. The signicant inhibitionowth could be observed in both oral administration andeal injection. FAK and phospho- FAK immunouores-g indicated that YZX lowered the expression of FAK andit the activation of FAK through treatment with a com-

DME25 and FAK inhibitor. Hence, YZX demonstratess a novel antimetastatic agent, targeting key traits inatic cascade and demonstrating efcacy using in vivo

Please cite

this

article in

press

as: Jian

g W

G,

et al.

Tissue

invasion

and

metastasis:

Molecu

lar, biological

and

clinical

persp

ectives. Sem

inCan

cer B

iol (2015),

http

://dx.d

oi.org/10.1016/j.semcan

cer.2015.03.008

ARTICLE IN PRESS

G M

odelYSC

BI-1183;

No.

of Pages

32

16

W.G

. Jiang

et al.

/ Sem

inars in

Cancer Biology

xxx (2015)

xxxxxx

Table 2Priority targets for tissue invasion and metastasis.

Priority targets fortissue invasion andmetastasis othercancer hallmarks

Upregulation ofE-cadherin

Promotion offormation oftight junctions(claudins, etc.)

Suppression ofsynthesis,secretion and/oractivity of theurokinaseplasminogenactivator (uPA)

Inhibition ofPI3K/AKTsignaling

Inhibition ofFAK signaling

Inhibition ofAP-1 activity

Inhibition ofNF-B

Suppression ofsynthesis,secretionand/or activityof MMP-9expression

Inactivation of-catenin/ZEB1signaling

Inhibition ofTGF-signaling

Genomic instability +[299]

0 0 +[300]

0 0 +[301303]

0 0 0

Sustained proliferativesignaling

+/[304,305]

+[306,307]

+[308,309]

+[310,311]

+[311,312]

+[313]

+[314316]

+[317,318]

+[319]

+[278,320]

Tumor-promotinginammation

+/[321,322]

+[323,324]

0 +[325327]

+[328,329]

+[330,331]

+[332,333]

+[334]

+[335]

+[336,337]

Evasion of anti-growthsignaling

+/[304,338340]

+[341,342]

0 +[343]

+[344,345]

+[346,347]

0 +[348,349]

+[350]

+/[255,351353]

Resistance to apoptosis +[354]

+[355]

+[356]

+[357]

+[358]

+[359]

+[360]

+[361]

+[362]

+[262]

Replicativeimmortality

0 0 [363]

+/[364366]

+[367]

[368]

+/[369371]

+[367]

0 [372,373]

Deregulatedmetabolism

+[374,375]

0 +[376]

+[377]

+[378]

0 +[379]

0 0 +[380]

Immune systemevasion

0 0 +[381]

+/[382,383]

0 0 +[384]

0 0 +[385]

Angiogenesis [386]

[387]

+[388]

+[389]

+[390,391]

+[392394]

+/[395,396]

+[397]

+[398]

+/[399]

Tumormicroenvironment

+/[400,401]

+[402]

+[309]

+[403]

+[404,405]

+[406,407]

+[408]

+/[409,410]

+[319]

+[411]

Key targets identied in tissue invasion and metastasis were also examined in the other hallmark areas of cancer. Targets relevant to other hallmarks are listed as complementary (+) if they display anti-carcinogenic effects,contrary () if they display pro-carcinogenic effects or controversial (+/) if they display both anti- and pro-carcinogenic affects, or identied as having no known relationship (0).

Please cite this article in press as: Jiang WG, et al. Tissue invasion and metastasis:Cancer Biol (2015), http://dx.doi.org/10.1016/j.semcancer.2015.03.008


Table

3Po

ssib

le

appro

aches

to

impac

t

prior

ity

targ

ets

for

tiss

ue

inva

sion

and

met

asta

sis.

Appro

aches

other

cance

r

hal

lmar

ksGam

ma

linol

enic

acid

Eico

sapen

tanoi

cac

id-(1

6)-d

-gl

uca

n(A

garicu

sbl

azei)

Grifo

lin

(Alb

atre

llus

con

uens

)

Cor

dyc

epin

(Cor

dyce

psm

ilita

ris)

Poly

sacc

har

ides

(Gan

oder

ma

luci

dum

)

Gan

oder

icac

ids

(Gan

oder

ma

luci

dum

)

Pach

ymic

acid

(Por

ia

coco

s)Si

libi

nin

5,6-

dih

ydro

-4H

-pyr

rolo

[1,2

-b]-

pyr

azol

es

Gen

omic

inst

ability

0

0

0

0

0

0

0

0

+ [412

414

]0

Sust

ained

pro

life

rative

sign

alin

g+ [4

15,4

16]

+ [417

]+ [4

18]

+ [419

]+ [4

204

23]

+/

[424

,425

]+ [4

26]

+ [427

]+ [4

284

31]

0

Tum

or-p

rom

otin

g

inam

mat

ion

0

+ [432

,433

]+ [4

34]

0

+ [435

,436

]+ [4

37,4

38]

+ [439

]+ [4

40]

+ [441

]0

Evas

ion

of

anti-g

row

th

sign

alin

g+

[442

,443

]+ [4

444

46]

0

+ [447

]+ [4

22,4

48]

+ [449

451

]+ [4

39,4

52]

+ [427

]+ [4

534

55]

0

Res

ista

nce

to

apop

tosis

+ [456

]+ [4

57]

+ [458

]+ [4

19]

+ [459

]+ [4

60]

+ [461

]+ [4

62]

+ [463

]0

Rep

lica

tive

imm

orta

lity

0

+ [464

,465

]0

0

0

0

0

0

+ [466

468

]0

Der

egula

ted

met

abol

ism

+ [469

]0

+ [470

]+ [4

71]

+ [472

]+ [4

73]

0

0 + [4

29,4

74]

0

Imm

une

syst

em

evas

ion

0

0

0

0

0

+ [475

]0

0 + [4

76]

0

Angi

ogen

esis

+ [416

]+ [4

77]

0

0

+ [478

]+ [4

79]

+ [439

]0

+ [230

]0

Tum

or

micro

envi

ronm

ent

+ [480

Documents

Tissue Invasion and Metastasis - Molecular, Biological and Clinical - Review