mechanisms of lymphatic metastasis in human colorectal adenocarcinoma

Journal of PathologyJ Pathol 2009; 217: 608–619Published online 13 January 2009 in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/path.2517

Review Article

Mechanisms of lymphatic metastasis in human colorectaladenocarcinomaDaniel Royston and David G. Jackson*MRC Human Immunology Unit, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Headington, Oxford OX3 9DS, UK

*Correspondence to:David G. Jackson, MRC HumanImmunology Unit, WeatherallInstitute of Molecular Medicine,John Radcliffe Hospital,Headington, Oxford OX39DS, UK.E-mail: [email protected]

No conflicts of interest weredeclared.

Received: 26 September 2008Revised: 12 December 2008Accepted: 25 December 2008

AbstractThe invasion of lymphatic vessels by colorectal cancer (CRC) and its subsequent spreadto draining lymph nodes is a key determinant of prognosis in this common and frequentlyfatal malignancy. Although tumoural lymphangiogenesis is assumed to contribute to thisprocess, review of the current literature fails to support any notion of a simple correlationbetween lymphatic vessel density and CRC metastasis. Furthermore, attempts to correlatethe expression of various lymphangiogenic growth factors, most notably VEGF-C andVEGF-D, with the lymphatic metastasis of CRC have provided contradictory results. Recentevidence from animal and human models of tumour metastasis suggests that complexfunctional and biochemical interactions between the microvasculature of tumours and othercell types within the tumour microenvironment may play a pivotal role in the behaviourof commonly metastasizing tumours. Indeed, previous insights into tumoural blood vesselshave provided candidate markers of tumoural angiogenesis that are currently the subjectof intense investigation as future therapeutic targets. In this review article we survey thecurrent evidence relating lymphangiogenesis and lymphangiogenic growth factor productionto metastasis by CRC, and attempt to provide some insight into the apparent discrepancieswithin the literature. In particular, we also discuss some new and provocative insights intothe properties of tumoural lymphatics suggesting that they have specific expression profilesdistinct from those of normal lymphatic vessels and that appear to promote metastasis.These findings raise the exciting prospect of future biomarkers of lymphatic metastasis andidentify potential targets for new generation anti-tumour therapies.Copyright 2009 Pathological Society of Great Britain and Ireland. Published by JohnWiley & Sons, Ltd.

Keywords: colorectal cancer; lymphatics; lymphangiogenesis; metastasis; lymph node;immunohistochemistry; ESAM

Introduction: predicting metastasis incolorectal cancer — what’s the problem?

Colorectal adenocarcinoma (CRC) is a common humanmalignancy with an annual incidence of over 1 millioncases worldwide [1]. It is responsible for approxi-mately 10% of all cancer-related deaths in Westerncountries and accounts for a significant proportion oftotal healthcare expenditure [2]. This expenditure isset to increase significantly in countries such as theUK, where all adults aged 60–69 will soon be invitedto participate in a nationwide screening programmefor CRC.

Although one of the better characterized tumourentities with regard to genetic progression, the out-come of patients with CRC is still largely determinedby a rather crude set of prognostic indicators. Of these,the presence of lymphatic metastasis to regional lymphnodes (LN) is particularly important. Indeed, both ofthe commonly used staging systems for CRC (Dukesand TNM) depend upon an assessment of the extent of

tumour spread through the bowel wall and the numberof lymph nodes containing metastatic tumour deposits(the Dukes staging system for CRC is illustrated inFigure 1). Early stages of tumour growth with no evi-dence of LN spread (Dukes A or B; pT1-3, N0) havean excellent post-operative prognosis and a 5 yearsurvival rate of 80–90% [3]. By contrast, tumoursbreaching the bowel wall serosa or invading adjacentstructures (pT4) or involving regional LNs (Dukes C)have a 5-year survival rate of only 25–60% [4]. Inaddition to its value as a prognostic indicator, the nodalstatus of CRC is of critical importance to oncologistsseeking to identify patients likely to benefit from adju-vant or neo-adjuvant chemo/radiotherapy [5].

Despite recent advances in radiological imaging,the most reliable way of detecting LN metastasisin CRC remains the careful microscopic assessmentof regional LNs harvested from resection specimens.However, the prognostic value of LN assessment isrestricted by the variable yield of nodes harvestedfrom surgical specimens [3] and may be adversely

Copyright 2009 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.www.pathsoc.org.uk

Mechanisms of lymphatic metastasis in human colorectal adenocarcinoma 609

Figure 1. Dukes staging of CRC. Dukes A tumours are confined to the bowel wall and show no lymph node spread. Dukes Btumours invade beyond the muscularis propria but show no lymph node spread; they may involve the free peritonealized surfaceof the bowel wall or other adjacent organs. Dukes C1 tumours spread to regional lymph nodes but show no involvement of theapical node. Dukes C2 tumours spread to the apical node

affected by the current recommendations to examinemicroscopically only one slide from each node [6].Consequently, accurate assessment of the risk oflymphatic metastasis is not always possible. Addedto this is the increasing use of conservative localexcision techniques for low rectal carcinomas [1].Unlike traditional resections, local excision techniquesprovide no regional LNs for histological assessment,and important management decisions have to be madewithout histological evidence of the tumour’s nodalstatus. In light of these considerations, there is a clearneed to identify CRC patients who have developed, orare at high risk of developing, lymphatic metastasis ina manner that is more accurate and reliable than thetraditional pathological assessment of excised LNs.

Lymphangiogenesis and its predictivevalue in CRC

Traditionally, lymphatic metastasis was considered tobe an essentially passive process in which malignantcells shed by solid tumours gain ready access to localpre-existing lymphatics due to their thin walls andincomplete basement membranes. This simple mecha-nistic explanation of tumour cell metastasis has beenradically revised in light of recent insights into tumourlymphatic biology made possible with the discovery ofseveral markers with specificity for the lymphatic vas-culature. Of these, VEGFR-3, PROX-1, podoplanin,

D6 and LYVE-1 have proved to be of particularvalue in the investigation of tumour lymphatics (seeTable 1 and references therein), most notably for theassessment of lymph vessel densities. Furthermore,the concurrent discovery of a growing repertoire oflymphangiogenic growth factors has provided someinsights into the molecular processes affecting lym-phatic vessel development and proliferation withintumour tissues. Consequently, numerous investigatorshave attempted to relate the density of lymphatic ves-sels in tumours and/or their expression of lymphan-giogenic growth factors to lymphatic metastasis. Thebackground to these closely related fields of investi-gation will briefly be considered prior to discussion oftheir specific relevance to CRC.

Increased lymphatic vessel densities intumours — are they significant?

Using the newly available lymphatic-specific mark-ers, several important preliminary studies examin-ing the artificial induction of lymphangiogenesis inxenotransplanted human tumours revealed intratu-moural and peritumoural lymphatic proliferation tobe strongly associated with lymphatic tumour spread[7–9]. Indeed, it was anticipated that tumoural lym-phangiogenesis leading to an increased lymphatic ves-sel density would provide a convenient and reliablemeans of determining the risk of lymphatic metas-tasis by solid tumours and act as a simple target

J Pathol 2009; 217: 608–619 DOI: 10.1002/pathCopyright 2009 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

610 D Royston et al

Table 1. Lymphatic endothelial markers commonly used in theinvestigation of tumour lymphatics

Lymphaticendothelialmarker Comments

VEGFR-3 Tyrosine kinase receptor for VEGF-C and VEGF-D[100, 101]Expressed by adult lymphatic vesselsRe-expression noted in blood capillaries in somevascular and breast tumours [102]

PROX-1 Homeobox domain transcription factor [103]Maintains differentiation of lymphatic endothelium inadult tissue [104]Required for the maintenance of embryonic precursorLEC and differentiation towards the lymphaticphenotype [105]

Podoplanin Cell surface sialylated glycoprotein [106]Precise function unknown

D6 Endocytic 7-TM receptor for inflammatory CCchemokines [107]

LYVE-1 Primary lymphatic endothelial receptor for hyaluronan[108]Precise function unknown

of future anti-tumour therapies. However, subsequentstudies in various human tumours failed to support thenotion that a simple relationship exists between tumourlymphatic vessel density (LVD) and clinicopatholog-ical variables such as LN metastasis and disease-free survival. Indeed, there is now growing evidencethat the pattern of tumoural lymphangiogenesis variesbetween tumour histological type and anatomical site.In several common human tumours, such as cuta-neous melanoma [10–12], head and neck squamouscell carcinoma (HNSCC) [13–15], transitional cellcarcinoma of the bladder [16,17] and non-small celllung cancer [18], intratumoural lymphangiogenesiscan be readily appreciated and has been shown tobe of some prognostic significance. By contrast, inbreast [19,20], cervical [21–23] and prostate carci-noma [24–28], tumours that also readily spread viathe lymphatics, there is little evidence of significantintratumoural lymphangiogenesis, with most prolifer-ating vessels lying within the peritumoural tissues. Ofnote, despite HNSCC and cervical cancers displayingsome histogenetic similarity, both arising from squa-mous epithelium, the patterns of lymphangiogenesisin these tumours are quite different. Although the rea-sons for this are unclear, these findings likely reflectthe importance of the local microenvironment in thedevelopment of primary tumours.

To date, a clear understanding of the roles playedby intra- and peritumoural lymphatics in these humantumours remains elusive and somewhat controversial,with several commentators even questioning the func-tional activity of intratumoural lymphatics due to thehigh compressive forces existing within solid tumours[29–32]. Taken together, these findings have proveddifficult to reconcile with a simple, universal model oftumour cell metastasis via the lymphatic vasculatureof primary tumours.

Lymphatic vessel densities in CRC

It is only relatively recently that the role of lymphan-giogenesis and the significance of LVD in human CRChave been examined in any detail. Several immuno-histochemical studies have revealed significant num-bers of peritumoural lymphatic vessels associated withprimary human CRCs [33–37] and CRC cell linesorthotopically implanted into mice [38]. By contrast,intratumoural lymphatics in CRC appear scarce andtypically have a distorted and compressed morphol-ogy [33,35,37], although the evidence for the clinicalsignificance of these vessels is contradictory. In atleast two studies, increased peritumoural LVD wasassociated with lymphatic vessel invasion and lymphnode metastasis [33,37]. However, contradictory evi-dence also exists, with at least one report show-ing no association between the extent of lymphan-giogenesis in CRC and clinicopathological variablesincluding patient outcome [35]. This same study alsorevealed an increase in the proliferative index of bothintra- and peritumoural lymphatics in comparison to‘normal’ tissues by immunohistochemical detectionof Ki-67, although no correlation with any clinico-pathological variable was established. Taken together,these findings support the notion of lymphangiogen-esis in CRC, but with vessel growth appearing to beprimarily peritumoural. However, the clinical signif-icance of lymphangiogenesis in CRC remains uncer-tain.

A note of caution in interpreting these studiesderives from the tendency of individual researchersto use different methods of identifying and scoringlymphatic vessels within and around CRC tissues. Inparticular, various lymphatic endothelial cell (LEC)markers with differing specificities have often beenused and several of the methodologies for countingpositively stained vessels fail to demonstrate clear evi-dence of low intra- and inter-observer variability (seeTable 2 for a summary of the lymphatic markers andvessel-counting strategies adopted in the literature). Inlight of such considerations, an international consen-sus report on the assessment of lymphangiogenesis hasrecently been published, which provides a detailed dis-cussion of methodologies used for the quantification oflymphangiogenesis in solid human tumours [39]. Theauthors recommend the use of at least one validatedlymphatic marker, although two are preferable in lightof the documented lack of absolute specificity for anysingle lymphatic marker. It is also recommended thatvessels be quantified using the Chalkley point gratic-ule method on manually selected vascular hot spotsin viable tumour and adjacent stromal tissues. Fur-thermore, the quantification of lymphatic endothelialcell proliferation should be performed by counting thenumber of proliferating lymphatic endothelial cells,using a reliable proliferation marker such as Ki-67,and all scores should be independently assessed by atleast two investigators.



Table 2. Lymphatic vessel densities and lymphangiogenic growth factor production and their relationship to colorectal cancer

Tumour stageNo. ofcases

Lymphaticmarker

Method ofassessing LVD

Method of assessing lymphangiogenicgrowth factor production

No. of LNharvested

Early stage (T1) 87 [33] Podoplanin LV counts from multiple‘hotspots’

ND ND

Mixed stage (T2–4) 24 [34] 5′-NA LV counts from three‘hotspots’

IHC NDHigh VEGF-C = cytoplasmic staining of ≥ 30%of tumour cellsLow VEGF-C = cytoplasmic staining of ≤ 30%of tumour cells

Mixed stage(Dukes A–D)

(Dukes A–D)

(Dukes A–C)

(Dukes A–C)

(Dukes A–C)

64 [35] Podoplanin LV counts from three‘hotspots’

ND ND

ND 15 [36] LYVE-1 LV counts from 10 ‘hotspots’+ qRT–PCR

ND ND

ND 268 [37] Podoplanin LV counts from three‘hotspots’

ND ND

Advanced stage(Dukes B–D)

152 [62] ND ND IHC + ISH for VEGF-C NDPositive VEGF-C = cytoplasmic staining of≥ 10% of tumour cells at deepest invasive site,central and superf icial tumour

Mixed stage 56 [63] 5′-NA LV counts from three

‘hotspots’

IHC ND

Positive VEGF-C = cytoplasmic staining of≥ 10% of tumour cells

Mixed stage (T1–4) 53 [64] ND ND IHC for VEGF-C ND82 [65] Positive VEGF-C = cytoplasmic staining of

≥ 10% of tumour cells99 [67]Early stage (T1) 221 [68] ND ND IHC ND

Positive VEGF-C = cytoplasmic staining of≥ 20% of tumour cells


152 [69] ND ND IHC NDPositive VEGF-C = cytoplasmic staining of≥ 10% of tumour cells in deepest invasive siteand superf icial areas

ND 28 [70] LYVE-1 for VEGF-C/D NDPROX-1Podoplanin5′-NA

Mixed stage 71 [71] ND ND RPA + RT–PCR + IHC for VEGF-A/C/D ND

Mixed stage 70 [72] VEGFR-3 ND IHC + RT–PCR for VEGF-A/C/D NDPositive VEGF-A/C/D = cytoplasmic staining of≥ 20% of tumour cells

Mixed stage (T1–4) 83 [73] ND ND IHC for VEGF-D NDPositive VEGF-D = cytoplasmic staining of≥ 10% of tumour cells

Mixed stage 84 [74] VEGFR-3 LV counts from five‘hotspots’

IHC for VEGF-D NDNegative staining scored 0Weak/limited staining scored 1Moderate/widespread staining scored 2Strong widespread staining scored 3


100 [75] ND ND IHC for bFGF NDProportion of tumour cells and their stainingintensity scored 0–3


101 [76] ND ND IHC for bFGF NDProportion of tumour cells and their stainingintensity at deepest invasive margin scored 0–3

ND, not determined; IHC, immunohistochemistry; 5′NA, 5′-nucleotidase; ISH, in situ mRNA hybridization analysis; RPA, ribonuclease protectionassay

qRT–PCR

qRT–PCR qRT–PCR

Lymphangiogenic growth factors — an updateBy far the most extensively studied lymphangio-genic factors are members of the vascular endothelialgrowth factor (VEGF) family of molecules, in par-ticular VEGF-C and VEGF-D. These secreted glyco-proteins largely signal via the cell surface tyrosine

kinase receptor VEGFR-3/Flt4 present on the sur-face of lymphatic endothelial cells (LEC), and resultin lymphatic vessel proliferation in vitro and in vivo[40–43]. Indeed, in several animal tumour modelsthe induction of lymphangiogenesis by engineeredover-expression of VEGF-C/D is sufficient to induce


612 D Royston et al

tumour invasion and lymph node metastasis [7–9,44].Furthermore, this process of VEGF-C/D-driven metas-tasis can be inhibited by administration of solubleVEGFR-3 [45] or monoclonal antibodies to VEGF-D[9] and VEGFR-3 [46]. Despite its initial recognitionas a potent factor in haemangiogenesis, VEGF-A hasalso been shown to promote lymphangiogenesis. Itcan act both indirectly via the recruitment of VEGF-C/D-producing inflammatory cells [47] and directlyvia its receptor (VEGFR-2) [48], which is expressedon LECs. VEGF-A has been shown to induce intra-and peritumoural lymphangiogenesis in transgenic andxenographic mouse tumour models and to promotelymph node metastasis [49,50].

In addition to members of the VEGF family ofmolecules, a large number of other growth factorsare now known to play important roles in the devel-opment, proliferation, branching and maintenanceof lymphatic vessels. These include platelet-derivedgrowth factor-BB (PDGF-BB) [51], fibroblast growthfactor 2 (FGF2) [10,52], angiopoietin 1 and 2 (Ang-1and Ang-2) [53–55], hepatocyte growth factor (HGF)[56,57] and insulin-like growth factor (IGF-1 and IGF-2) [58]. These molecules appear to have interdepen-dent or collaborative roles in the establishment offunctional lymphatics, and their contribution to ourcurrent understanding of lymphangiogenesis has beenwell documented and extensively reviewed [59–61].

Lymphangiogenic growth factors in CRC

Various human and animal studies have attemptedto demonstrate an association between expressionof the VEGF family of molecules and clinical out-come in CRC. Significant VEGF-C expression hasbeen demonstrated immunohistochemically at the deepinvasive margin in almost half of resected CRCsand this has been correlated with various clinico-pathological variables, including lymphovascular inva-sion (LVI), LN metastasis and poor overall survival[62–69]. Importantly, the detection of VEGF-C incentral or superficial parts of colorectal tumours hasbeen shown to have no association with any clinico-pathological variables [62], raising doubts about thepotential prognostic value of these observations in rou-tine clinical practice, since these are the regions oftumour tissue most likely to be sampled by endoscopicbiopsy. Conflicting evidence regarding the potentialrole of VEGF-C as a useful prognostic marker inCRC has also been documented. One study showedno evidence of increased tumoural VEGF-C mRNAexpression in CRC [70], with at least two studiesshowing no association between VEGF-C expressionand LN metastasis [71,72]. Increased expression levelsof VEGF-D mRNA and protein have been demon-strated in CRC [70], with some studies showingan association between raised VEGF-D and adverseclinicopathological variables, including LN metastasisand reduced patient survival [66,73]. However, othergroups have shown VEGF-D mRNA expression to be

significantly lower in CRC tissues when comparedto normal mucosa, with no correlation to lymphaticmetastasis [65,71,72]. In one study, VEGF-D expres-sion was even shown by multivariate analysis to bean independent prognostic factor for both disease-freeand overall survival [74]. VEGF-A mRNA and pro-tein expression has been observed in CRC [71], withincreased levels shown to correlate with lymphaticmetastasis and poor prognosis by at least two groups[69,72].

Relatively little is known about the expression andclinical significance of other lymphangiogenic factorsin CRC. However, a correlation between tumour bFGFexpression, lymphatic invasion and LN metastasis hasbeen reported [75] while Ang-2 over-expression hasbeen shown to correlate with lymphatic invasion, LNmetastasis and poor prognosis [76].

Taken together, there is no clear consensus in the lit-erature regarding the expression of lymphangiogenicgrowth factors in CRC and their association with clin-icopathological variables such as LN metastasis andpatient survival. However, interpreting these findingsis complicated by the differing methods and techniquesemployed for quantifying lymphangiogenic growthfactors in human and animal tissues (see Table 2).These differences may partly explain some of thecontradictory results in the literature. In particular,the assessment of immunohistochemical staining intumour tissues is highly subjective and a lack ofconsistency in the scoring methods employed makescomparisons between different studies difficult. Fur-thermore, the observed heterogeneity of tumour cellimmunoreactivity for various growth factors highlightsthe importance of discriminating superficial and cen-tral portions of tumour from the invasive margin. Dif-ficulties also arise when trying to interpret the findingsof studies examining the mRNA expression of variouslymphangiogenic molecules. By necessity, determina-tion of mRNA expression involves calculating valuesfor tumour tissue as a whole, with no discriminationbetween tumour cells, stroma and benign peritumouraltissues. Consequently, important differences in proteinexpression between various components in the tumourmicroenvironment may be missed and total mRNAlevels may not correlate with the actual protein expres-sion levels observed within tumours. Indeed, VEGF-Cand VEGF-D expression has been described within theinflammatory infiltrates that frequently accompany theinvasive margins of some aggressive human tumours.In particular, these lymphangiogenic growth factorsappear to be expressed by distinct populations oftumour-associated macrophages (TAMs) whose den-sity within the inflammatory infiltrate is associatedwith lymphatic vessel invasion [22,77].

‘Non-lymphangiogenic’ mechanisms oflymphatic metastasis

As outlined above, there is little to suggest that asimple relationship exists between the density and



pattern of lymphatic vessel proliferation in CRCand LN metastasis. Furthermore, quantification ofvarious lymphatic growth factors within CRC suggeststhat the presence of lymphovascular invasion andLN metastasis is not simply a result of elevatedgrowth factor secretion and its direct lymphangiogeniceffects. In light of these observations, various othermechanisms have been invoked to explain the processof tumour cell infiltration and spread via the lymphaticvasculature (summarized in Figure 2).

In vivo studies have shown that VEGF-C can sig-nificantly increase the contraction frequency, end dias-tolic diameter, stroke volume index and pump flowindex of collecting lymphatics [78] and induce hyper-plasia of peritumoural lymphatics with an associatedincrease in the rate of lymph node metastasis [44,79].These processes were shown to be a direct conse-quence of altered lymphatic function, rather than achange in tissue haemodynamics and were exclusivelymediated via activation of VEGFR-3. The potential forsuch mechanisms to increase the lymphatic drainage ofCRC tissues and thereby contribute to LN metastasisremains to be elucidated.

The study of tumoural lymphangiogenesis haslargely been confined to the primary site of tumourgrowth and adjacent tissues. However, regional lymphnodes, in particular the sentinel nodes that draintumours, can also display lymphangiogenesis. Curi-ously, in murine models of chemically inducedskin carcinogenesis which over-express VEGF-A andVEGF-C, sentinel node lymphangiogenesis has evenbeen noted prior to the arrival of metastasizing tumourcells [49,80]. Furthermore, murine tumour modelsof nasopharyngeal carcinoma [81] and melanoma

[82] have revealed enlarged lymphatic sinuses, withconcomitant increases in the lymphatic drainage ofnodes displaying no evidence of tumour cell seeding.Such insights have led to speculation that tumourswith metastatic potential may condition downstreamlymph nodes via lymph-borne signals in order to opti-mize metastatic tumour cell seeding and develop spe-cific pathways of increased tumour lymph drainage(reviewed in [60]). The extent to which CRCs mayinduce downstream proliferative changes in the lym-phatics of draining lymph nodes, or modify the volumeand character of their lymphatic drainage, remains tobe investigated.

Finally, cell-specific interactions between LECs andtumour cells may play an important role in the pro-cess of lymphatic vessel invasion. Indeed, it is wellestablished that tumour LEC chemotaxis and prolif-eration can be driven by tumour-derived VEGF-C,the so-called ‘chemotaxis–lymphangiogenesis hypoth-esis’ [83,84]. However, recent work using both LECmigration assays and an in vivo murine melanomamodel suggests that LECs secrete chemotactic agentsthat may directly attract cancer cells. Conditionedmedium from LECs was shown to attract metastasizingmelanoma cells expressing CCR-7 receptors for thechemokine CCL21/SLC, a process inhibited in vivoby neutralizing antibodies to CCL21 [84]. Over-expression of CXCR-4 in an oral squamous cell car-cinoma has also been shown to promote lymph nodemetastasis in an orthotopic mouse model [85], whileblocking of the CXCL12(SDF-1)–CXCR-4 interactionin a murine breast carcinoma model was shown tosignificantly inhibit lymphatic metastasis by tumourcells [86]. In the context of CRC, one animal study

Figure 2. Simplified overview of the proposed mechanisms of LN metastasis. Various models have been developed to explainthe infiltration of lymphatics by tumour cells and their metastasis to regional LNs. Tumours may secrete lymphangiogenic growthfactors (e.g. VEGF-C/D) which increase the total lymphatic vessel surface area available for invasion. Tumour-secreted factorscould also increase the pumping action of draining afferent lymphatic vessels, or induce LN lymphangiogenesis in preparationfor metastasis. Alternatively, tumour cells expressing receptors (e.g. CCR-7, CXCR-4, CXCR-3) for specific LEC-secretedchemokines (e.g. CCL21/SLC, CXCL12/SDF-1) may approach and invade pre-existing or newly dividing tissue lymphatics. Finally,recent evidence suggests that tumour LECs may display an altered phenotype that is exploited by tumour cells to promote orenhance lymphatic metastasis


614 D Royston et al

has demonstrated that human cancer cells express-ing CXCR-3 display more rapid metastasis and nodalexpansion when compared to non-CXCR-3 express-ing tumour cells. Indeed, elevated CXCR-3 expressionin human CRC specimens was seen to be associatedwith poor prognosis [87]. Retrospective assessmentof archival CRC specimens has also shown an asso-ciation between elevated levels of CCR-7/CXCR-4and lymph node status and reduced patient survival[88]. These studies suggest that, while the produc-tion of lymphangiogenic growth factors by tumourcells may attract LECs, chemokines secreted by lym-phatic vessels may attract the movement of tumourcells expressing cognate chemokine receptors andthereby facilitate lymphatic metastasis. The extent towhich the chemokine-induced reciprocal attraction oftumour cells and LECs actually drives tumour cellmetastasis in vivo is currently unknown. However,there is growing support for the concept that tumourlymphatics may display specific functional propertiesthat directly contribute to tumour cell infiltration andmetastasis.

An alternative mechanism for lymphaticmetastasis

The notion that qualitative changes in the lymphaticvessels of tumours may be as important as theirtotal number and distribution is strongly supported byrecent studies of tumour blood vessels, in which RNAprofiling of haemovascular endothelium by sequen-tial analysis of gene expression (SAGE) identifiedalterations in gene expression between tumour andnormal blood vessels [89]. In turn, such studies identi-fied a number of tumour endothelial markers (TEMs)that are mechanistically implicated in tumour metas-tasis [90,91]. That tumoural lymphatics may also befunctionally distinct from those of normal tissueshas been supported by in vivo screening of phage-display libraries, which have identified peptides capa-ble of homing specifically to the lymphatics of certaintumour tissues [92]. However, until very recently nostudies have sought to identify specific differences inthe gene expression profiles of tumour and normallymphatic vessels. Consequently, there has been noprospect of identifying lymphatic markers analogousto the TEMs of the blood vasculature, or of unravel-ling the specific processes that might enable metastatictumour cells to gain access to the lymphatic compart-ment.

In response, our laboratory has recently under-taken the first GeneChip microarray and compar-ative immunohistochemical analysis of normal andtumour-associated LECs isolated from a congenicmouse tumour model — the lymph node metastasiz-ing T-241/VEGF-C fibrosarcoma [93]. RNA profil-ing revealed 792 transcripts with significantly alteredexpression (≥two-fold up-/down-regulation) includingthose that code for a variety of proteins involved in

components of endothelial cell junctions, subendothe-lial matrix and vessel growth/patterning. Importantly,the expression profile of metastatic tumour LECs wasfound to be distinct from that of non-metastatic tumourLECs, normal dermal LECs, mitogen (VEGF-C)-activated [94] or inflammation-activated LECs [95].Of particular interest, we identified three molecules:endothelial specific adhesion molecule (ESAM), anendothelial tight junction protein recently shown toregulate neutrophil transmigration across blood vesselendothelium [96]; Endoglin (TGFβ 1/3 receptor), acomponent of the transforming growth factor receptorcomplex associated with haemangiogenesis [97]; andthe angiogenesis-associated Leptin receptor (Leptin-R), whose up-regulation in the murine fibrosarcomamodel was similarly identified in the lymphatics ofseveral commonly metastasizing human malignancies.ESAM and Endoglin were both up-regulated in thelymphatics of CRC(Figure 3) in addition to HNSCC,ovarian carcinoma, lung adenocarcinoma and uterinemalignant mixed mullerian tumour (MMMT). Signifi-cantly, each of these tumours readily metastasizes viathe lymphatics to regional and distant LNs.

The common finding of similar changes in thephenotype of tumoural lymphatics derived from amurine tumour model and several human tumours indi-cates that alterations in the LEC transcription profileof tumours probably represent fundamental processesthat cross species barriers. Furthermore, the observedpattern of altered gene expression, which includesgenes involved in cell adhesion, transmigration andextracellular matrix production, strongly suggests thattumoural vessels have significantly altered cell junc-tions and an altered subendothelial matrix. Thesefindings tempt speculation that tumour LECs displayhighly specific functional alterations that directly con-tribute to the entry of metastasizing cells into thelymphatic compartment of tumours.

To examine the prognostic value of the altered phe-notype identified in the murine fibrosarcoma modeland validated in the human cancer tissues usingimmunofluoresence microscopy, the expression ofESAM and Endoglin in the lymphatics of two pan-els of HNSCC and colorectal carcinoma was exam-ined. Each panel contained an equal number (n = 28)of tumours matched by pathological stage or max-imum tumour dimension, with or without evidenceof LN metastasis. Quantitative analysis of immunos-tained vessels revealed a striking association betweenthe number of strongly stained Endoglin-positivelymphatic vessels within HNSCC tumours and thepresence of lymph node metastasis, with a greaterincidence of Endoglin-positive vessels observed intumours with LN metastasis. However, this associationwas less marked when weakly stained vessels wereassessed. Quantitative analysis of lymphatic ESAMstaining confirmed these findings, with a dramaticassociation between the incidence of ESAM-positive



Figure 3. Immunostaining for Endoglin and ESAM in colorectal adenocarcinoma. Representative sections of colorectaladenocarcinoma immunostained for LYVE-1/Alexa 568 (red) and either Endoglin or ESAM–Streptavidin Alexa 488 (green)to show expression in tumoural tissues compared to normal marginal tissues. Merged images include DAPI nuclear staining. Littleif any Endoglin and only trace levels of ESAM (arrows) are present in lymphatic vessel endothelium of normal tissue. In contrast,both Endoglin and ESAM are present in lymphatic vessel endothelium of colorectal tumour tissue

tumour vessels and the presence of lymph node metas-tasis in both HNSCC and CRC, with a greater inci-dence of ESAM-positive vessels observed in tumourswith LN metastasis (Figure 4).

In light of evidence that ESAM mediates neutrophiltransmigration of the blood vascular endothelium viaRho-mediated changes in the stability of tight junc-tions [96], it is tempting to speculate that similar

Figure 4. Association between ESAM-positive lymphaticvessel number and lymph node metastasis in colorectaladenocarcinoma. Scatter plot showing individual counts fortotal ESAM-positive lymphatic vessel numbers within thetumour and normal margin in 28 cases of colorectaladenocarcinoma categorized according to the presence orabsence of demonstrable LN metastases, as assessed by dualLYVE-1 immunofluorescence microscopy. Values shown aremean ± SEM. Reproduced with permission from Clasper et al.[93]

alterations in the tight junctions of tumour LECs maybe exploited by tumour cells during their invasionof lymphatic vessels. The observation that Endoglinis a marker of proliferating blood vessels [97] andan independent predictor of metastasis in severalhuman cancers [98] raises the interesting prospect thatthis protein may have related functions in the bloodand lymphatic vessels of tumours. The expression ofEndoglin by tumour LECs should also caution the useof this molecule as a specific marker of blood vascu-lar endothelium. Finally, the expression of Leptin-R ontumour lymphatics is a curious and potentially interest-ing finding. This receptor for the endocrine hormoneleptin, produced by adipose tissue and involved in thehypothalamic control of total body weight, has beenshown to induce the formation of vascular fenestra-tions on blood microvessels and increase their perme-ability [99]. This raises further speculation about thepossible significance of fat metabolism and obesity inthe context of malignancy.

In summary, based upon these findings, we pro-pose a new hypothesis in which the lymphatic vesselsof tumours display a degree of phenotypic plasticity,induced by mechanisms that remain to be elucidated,that facilitates tumour cell entry to the lymphaticsand thereby contributes to lymphatic tumour spread.Such qualitative changes in the nature of tumour lym-phatics would at least partially explain some of theapparent inconsistencies in the literature attempting topredict lymphatic vessel invasion solely on the basisof lymph vessel density or lymphangiogenic growthfactor secretion. The detailed characterization of suchqualitative changes may find application in the devel-opment of better prognostic indicators in cancers suchas CRC. In turn, this may contribute to the develop-ment of improved management strategies employed totreat this common and important malignancy.


616 D Royston et al

Conclusion

The lymphatic invasion and LN metastasis of CRCrepresents a critical phase in the natural biology ofCRC, and is a key prognostic factor for patients withthis common malignancy. The accurate assessment ofLN metastasis in patients newly diagnosed with CRCis critical if appropriate therapeutic regimes are to befollowed.

Notwithstanding several human and animal studiesof lymphangiogenesis and lymphangiogenic growthfactor production by CRC tissues, many fundamen-tal questions concerning the mechanisms by whichtumour cells gain entry to the lymphatic compartmentremain unanswered. There is now growing evidencethat specific, qualitative changes in the lymphatic ves-sels of tumours may be at least as important as theirnumber and distribution within tumour tissues. Byidentifying and fully characterizing the functional andbiochemical nature of these changes it may be pos-sible to gain a clearer understanding of how tumoursmetastasize via the lymphatics. Furthermore, it maybe possible to develop more accurate markers of dis-ease progression in CRC and other tumours, and toenlighten the search for future anti-tumour therapiesthat specifically target tumoural lymphatics.

Acknowledgements

This work was supported by Cancer Research UK (GrantNo. C581, to DGJ) and the National Coordinating Centre forResearch Capacity Development Fellowship scheme (to DR).

References

1. Jankowski J, Sampliner R, Kerr D, Fong Y (eds). Gastroin-testinal Oncology: A Critical Multidisciplinary Team Approach.Blackwell: Oxford, 2008.

2. Kumar V, Fausto N, Abbas A. Robbins & Cotran PathologicBasis of Disease, 7th edn. Saunders: Philadelphia, PA, 2004.

3. Swanson RS, Compton CC, Stewart AK, Bland KI. The progno-sis of T3N0 colon cancer is dependent on the number of lymphnodes examined. Ann Surg Oncol 2003;10:65–71.

4. Compton CC, Greene FL. The staging of colorectal cancer: 2004and beyond. CA Cancer J Clin 2004;54:295–308.

5. Sezeur A, Chatelet FP, Cywiner C, de Labriolle-Vaylet C,Chastang C, Billotey C, et al. Pathology underrates colon cancerextranodal and nodal metastases; ex vivo radioimmunodetectionhelps staging. Clin Cancer Res 2007;13:s 5592–5597.

6. Williams GT, Quirke P, Shepherd NA. Dataset for ColorectalCancer, 2nd edn. The Royal College of Pathologists: London,2007.

7. Skobe M, Hawighorst T, Jackson DG, Prevo R, Janes L, VelascoP, et al. Induction of tumor lymphangiogenesis by VEGF-Cpromotes breast cancer metastasis. Nat Med 2001;7:192–198.

8. Mandriota SJ, Jussila L, Jeltsch M, Compagni A, Baetens D,Prevo R, et al. Vascular endothelial growth factor-C-mediatedlymphangiogenesis promotes tumour metastasis. EMBO J2001;20:672–682.

9. Stacker SA, Caesar C, Baldwin ME, Thornton GE, Williams RA,Prevo R, et al. VEGF-D promotes the metastatic spread of tumorcells via the lymphatics. Nat Med 2001;7:186–191.

10. Straume O, Jackson DG, Akslen LA. Independent prognosticimpact of lymphatic vessel density and presence of low-grade

lymphangiogenesis in cutaneous melanoma. Clin Cancer Res2003;9:250–256.

11. Shields JD, Borsetti M, Rigby H, Harper SJ, Mortimer PS,Levick JR, et al. Lymphatic density and metastatic spread inhuman malignant melanoma. Br J Cancer 2004;90:693–700.

12. Dadras SS, Paul T, Bertoncini J, Brown LF, Muzikansky A,Jackson DG, et al. Tumor lymphangiogenesis: a novel prognosticindicator for cutaneous melanoma metastasis and survival. Am JPathol 2003;162:1951–1960.

13. Beasley NJ, Prevo R, Banerji S, Leek RD, Moore J, vanTrappen P, et al. Intratumoral lymphangiogenesis and lymphnode metastasis in head and neck cancer. Cancer Res2002;62:1315–1320.

14. Kyzas PA, Geleff S, Batistatou A, Agnantis NJ, Stefanou D.Evidence for lymphangiogenesis and its prognostic implicationsin head and neck squamous cell carcinoma. J Pathol2005;206:170–177.

15. Franchi A, Gallo O, Massi D, Baroni G, Santucci M. Tumorlymphangiogenesis in head and neck squamous cell carcinoma:a morphometric study with clinical correlations. Cancer2004;101:973–978.

16. Fernandez MI, Bolenz C, Trojan L, Steidler A, Weiss C, AlkenP, et al. Prognostic implications of lymphangiogenesis in muscle-invasive transitional cell carcinoma of the bladder. Eur Urol2008;53:571–578.

17. Miyata Y, Kanda S, Ohba K, Nomata K, Hayashida Y, Eguchi J,et al. Lymphangiogenesis and angiogenesis in bladder cancer:prognostic implications and regulation by vascular endothelialgrowth factors A, C, and D. Clin Cancer Res 2006;12:800–806.

18. Renyi-Vamos F, Tovari J, Fillinger J, Timar J, Paku S, Kenessey I,et al. Lymphangiogenesis correlates with lymph node metastasis,prognosis, and angiogenic phenotype in human non-small celllung cancer. Clin Cancer Res 2005;11:7344–7353.

19. Bono P, Wasenius VM, Heikkila P, Lundin J, Jackson DG,Joensuu H. High LYVE-1-positive lymphatic vessel numbers areassociated with poor outcome in breast cancer. Clin Cancer Res2004;10:7144–7149.

20. Williams CS, Leek RD, Robson AM, Banerji S, Prevo R, Har-ris AL, et al. Absence of lymphangiogenesis and intratumourallymph vessels in human metastatic breast cancer. J Pathol2003;200:195–206.

21. Van Trappen PO, Steele D, Lowe DG, Baithun S, Beasley N,Thiele W, et al. Expression of vascular endothelial growthfactor (VEGF)-C and VEGF-D, and their receptor VEGFR-3, during different stages of cervical carcinogenesis. J Pathol2003;201:544–554.

22. Schoppmann SF, Birner P, Stockl J, Kalt R, Ullrich R, Cau-cig C, et al. Tumor-associated macrophages express lymphaticendothelial growth factors and are related to peritumoral lym-phangiogenesis. Am J Pathol 2002;161:947–956.

23. Gombos Z, Xu X, Chu CS, Zhang PJ, Acs G. Peritumorallymphatic vessel density and vascular endothelial growth factor Cexpression in early-stage squamous cell carcinoma of the uterinecervix. Clin Cancer Res 2005;11:8364–8371.

24. Trojan L, Michel MS, Rensch F, Jackson DG, Alken P, Grob-holz R. Lymph and blood vessel architecture in benign andmalignant prostatic tissue: lack of lymphangiogenesis in prostatecarcinoma assessed with novel lymphatic marker lymphaticvessel endothelial hyaluronan receptor (LYVE-1). J Urol2004;172:103–107.

25. Trojan L, Rensch F, Voss M, Grobholz R, Weiss C, Jackson DG,et al. The role of the lymphatic system and its specific growthfactor, vascular endothelial growth factor C, for lymphogenicmetastasis in prostate cancer. BJU Int 2006;98:903–906.

26. Kuroda K, Horiguchi A, Asano T, Asano T, Hayakawa M.Prediction of lymphatic invasion by peritumoral lymphatic vesseldensity in prostate biopsy cores. Prostate 2008;68:1057–1063.

27. Cheng L, Bishop E, Zhou H, Maclennan GT, Lopez-Beltran A,Zhang S, et al. Lymphatic vessel density in radical prostatectomyspecimens. Hum Pathol 2008;39:610–615.



28. Roma AA, Magi-Galluzzi C, Kral MA, Jin TT, Klein EA,Zhou M. Peritumoral lymphatic invasion is associated withregional lymph node metastases in prostate adenocarcinoma. ModPathol 2006;19:392–398.

29. Leu AJ, Berk DA, Lymboussaki A, Alitalo K, Jain RK. Absenceof functional lymphatics within a murine sarcoma: a molecularand functional evaluation. Cancer Res 2000;60:4324–4327.

30. Jain RK, Fenton BT. Intratumoral lymphatic vessels: a caseof mistaken identity or malfunction? J Natl Cancer Inst2002;94:417–421.

31. Padera TP, Kadambi A, di Tomaso E, Carreira CM, Brown EB,Boucher Y, et al. Lymphatic metastasis in the absence offunctional intratumor lymphatics. Science 2002;296:1883–1886.

32. Padera TP, Stoll BR, Tooredman JB, Capen D, di Tomaso E,Jain RK. Pathology: cancer cells compress intratumour vessels.Nature 2004;427:695.

33. Liang P, Hong JW, Ubukata H, Liu HR, Watanabe Y, Katano M,et al. Increased density and diameter of lymphatic microvesselscorrelate with lymph node metastasis in early stage invasivecolorectal carcinoma. Virchows Arch 2006;448:570–575.

34. Ohno M, Nakamura T, Kunimoto Y, Nishimura K, Chung-Kang C, Kuroda Y. Lymphagenesis correlates with expression ofvascular endothelial growth factor-C in colorectal cancer. OncolRep 2002;10:939–943.

35. Omachi T, Kawai Y, Mizuno R, Nomiyama T, Miyagawa S,Ohhashi T, et al. Immunohistochemical demonstration of prolif-erating lymphatic vessels in colorectal carcinoma and its clinico-pathological significance. Cancer Lett 2007;246:167–172.

36. Gao F, Lu YM, Cao ML, Liu YW, He YQ, Wang Y. Expressionand quantification of LYVE-1 in human colorectal cancer. ClinExp Med 2006;6:65–71.

37. Kaneko I, Tanaka S, Oka S, Kawamura T, Hiyama T, Ito M,et al. Lymphatic vessel density at the site of deepest penetrationas a predictor of lymph node metastasis in submucosal colorectalcancer. Dis Colon Rectum 2007;50:13–21.

38. Onogawa S, Kitadai Y, Tanaka S, Kuwai T, Kuroda T, ChayamaK. Regulation of vascular endothelial growth factor (VEGF)-C and VEGF-D expression by the organ microenvironment inhuman colon carcinoma. Eur J Cancer 2004;40:1604–1609.

39. Van der Auwera I, Cao Y, Tille JC, Pepper MS, Jackson DG,Fox SB, et al. First international consensus on the methodologyof lymphangiogenesis quantification in solid human tumours. BrJ Cancer 2006;95:1611–1625.

40. Jeltsch M, Kaipainen A, Joukov V, Meng X, Lakso M, Rau-vala H, et al. Hyperplasia of lymphatic vessels in VEGF-C trans-genic mice. Science 1997;276:1423–1425.

41. Oh SJ, Jeltsch MM, Birkenhager R, McCarthy JE, Weich HA,Christ B, et al. VEGF and VEGF-C: specific induction ofangiogenesis and lymphangiogenesis in the differentiated avianchorioallantoic membrane. Dev Biol 1997;188:96–109.

42. Veikkola T, Jussila L, Makinen T, Karpanen T, Jeltsch M,Petrova TV, et al. Signalling via vascular endothelial growthfactor receptor-3 is sufficient for lymphangiogenesis in transgenicmice. EMBO J 2001;20:1223–1231.

43. Makinen T, Veikkola T, Mustjoki S, Karpanen T, Catimel B,Nice EC, et al. Isolated lymphatic endothelial cells transducegrowth, survival and migratory signals via the VEGF-C/Dreceptor VEGFR-3. EMBO J 2001;20:4762–4773.

44. He Y, Rajantie I, Pajusola K, Jeltsch M, Holopainen T, Yla-Herttuala S, et al. Vascular endothelial cell growth factor receptor3-mediated activation of lymphatic endothelium is crucial fortumor cell entry and spread via lymphatic vessels. Cancer Res2005;65:4739–4746.

45. He Y, Kozaki K, Karpanen T, Koshikawa K, Yla-Herttuala S,Takahashi T, et al. Suppression of tumor lymphangiogenesis andlymph node metastasis by blocking vascular endothelial growthfactor receptor 3 signaling. J Natl Cancer Inst 2002;94:819–825.

46. Roberts N, Kloos B, Cassella M, Podgrabinska S, Persaud K,Wu Y, et al. Inhibition of VEGFR-3 activation with theantagonistic antibody more potently suppresses lymph node anddistant metastases than inactivation of VEGFR-2. Cancer Res2006;66:2650–2657.

47. Cursiefen C, Chen L, Borges LP, Jackson D, Cao J, Radziejew-ski C, et al. VEGF-A stimulates lymphangiogenesis and heman-giogenesis in inflammatory neovascularization via macrophagerecruitment. J Clin Invest 2004;113:1040–1050.

48. Hong YK, Lange-Asschenfeldt B, Velasco P, Hirakawa S, Kun-stfeld R, Brown LF, et al. VEGF-A promotes tissue repair-associated lymphatic vessel formation via VEGFR-2 and theα1β1 and α2β1 integrins. FASEB J 2004;18:1111–1113.

49. Hirakawa S, Kodama S, Kunstfeld R, Kajiya K, Brown LF,Detmar M. VEGF-A induces tumor and sentinel lymph nodelymphangiogenesis and promotes lymphatic metastasis. J ExpMed 2005;201:1089–1099.

50. Bjorndahl MA, Cao R, Burton JB, Brakenhielm E, Religa P,Galter D, et al. Vascular endothelial growth factor-a promotesperitumoral lymphangiogenesis and lymphatic metastasis. CancerRes 2005;65:9261–9268.

51. Cao R, Bjorndahl MA, Religa P, Clasper S, Garvin S, Galter D,et al. PDGF-BB induces intratumoral lymphangiogenesis andpromotes lymphatic metastasis. Cancer Cell 2004;6:333–345.

52. Kubo H, Cao R, Brakenhielm E, Makinen T, Cao Y, Alitalo K.Blockade of vascular endothelial growth factor receptor-3 signaling inhibits fibroblast growth factor-2-induced lym-phangiogenesis in mouse cornea. Proc Natl Acad Sci USA2002;99:8868–8873.

53. Gale NW, Thurston G, Hackett SF, Renard R, Wang Q, McClainJ, et al. Angiopoietin-2 is required for postnatal angiogenesisand lymphatic patterning, and only the latter role is rescued byAngiopoietin-1. Dev Cell 2002;3:411–423.

54. Tammela T, Saaristo A, Lohela M, Morisada T, Tornberg J,Norrmen C, et al. Angiopoietin-1 promotes lymphatic sproutingand hyperplasia. Blood 2005;105:4642–4648.

55. Morisada T, Oike Y, Yamada Y, Urano T, Akao M, Kubota Y,et al. Angiopoietin-1 promotes LYVE-1-positive lymphatic vesselformation. Blood 2005;105:4649–4656.

56. Kajiya K, Hirakawa S, Ma B, Drinnenberg I, Detmar M. Hepa-tocyte growth factor promotes lymphatic vessel formation andfunction. EMBO J 2005;24:2885–2895.

57. Cao R, Bjorndahl MA, Gallego MI, Chen S, Religa P, Hansen AJ,et al. Hepatocyte growth factor is a lymphangiogenic factor withan indirect mechanism of action. Blood 2006;107:3531–3536.

58. Bjorndahl M, Cao R, Nissen LJ, Clasper S, Johnson LA, Xue Y,et al. Insulin-like growth factors 1 and 2 induce lymphangiogen-esis in vivo. Proc Natl Acad Sci USA 2005;102:15593–15598.

59. Sundlisaeter E, Dicko A, Sakariassen PO, Sondenaa K, Enger PO,Bjerkvig R. Lymphangiogenesis in colorectal cancer–prognosticand therapeutic aspects. Int J Cancer 2007;121:1401–1409.

60. Achen MG, Stacker SA. Molecular control of lymphaticmetastasis. Ann N Y Acad Sci 2008;1131:225–234.

61. Cao Y, Zhong W. Tumor-derived lymphangiogenic factors andlymphatic metastasis. Biomed Pharmacother 2007;61:534–539.

62. Furudoi A, Tanaka S, Haruma K, Kitadai Y, Yoshihara M,Chayama K, et al. Clinical significance of vascular endothelialgrowth factor C expression and angiogenesis at the deepestinvasive site of advanced colorectal carcinoma. Oncology2002;62:157–166.

63. Jia YT, Li ZX, He YT, Liang W, Yang HC, Ma HJ. Expressionof vascular endothelial growth factor-C and the relationshipbetween lymphangiogenesis and lymphatic metastasis incolorectal cancer. World J Gastroenterol 2004;10:3261–3263.

64. Kawakami M, Furuhata T, Kimura Y, Yamaguchi K, Hata F,Sasaki K, et al. Quantification of vascular endothelial growthfactor-C and its receptor-3 messenger RNA with real-timequantitative polymerase chain reaction as a predictor oflymph node metastasis in human colorectal cancer. Surgery2003;133:300–308.

65. Kawakami M, Furuhata T, Kimura Y, Yamaguchi K, Hata F,Sasaki K, et al. Expression analysis of vascular endothelialgrowth factors and their relationships to lymph nodemetastasis in human colorectal cancer. J Exp Clin Cancer Res2003;22:229–237.

66. Onogawa S, Kitadai Y, Tanaka S, Kuwai T, Kimura S, ChayamaK. Expression of VEGF-C and VEGF-D at the invasive edge


618 D Royston et al

correlates with lymph node metastasis and prognosis of patientswith colorectal carcinoma. Cancer Sci 2004;95:32–39.

67. Akagi K, Ikeda Y, Miyazaki M, Abe T, Kinoshita J, Mae-hara Y, et al. Vascular endothelial growth factor-C (VEGF-C)expression in human colorectal cancer tissues. Br J Cancer2000;83:887–891.

68. Maeda K, Yashiro M, Nishihara T, Nishiguchi Y, Sawai M,Uchima K, et al. Correlation between vascular endothelial growthfactor C expression and lymph node metastasis in T1 carcinomaof the colon and rectum. Surg Today 2003;33:736–739.

69. Kaio E, Tanaka S, Kitadai Y, Sumii M, Yoshihara M, Haruma K,et al. Clinical significance of angiogenic factor expression at thedeepest invasive site of advanced colorectal carcinoma. Oncology2003;64:61–73.

70. Parr C, Jiang WG. Quantitative analysis of lymphangio-genic markers in human colorectal cancer. Int J Oncol2003;23:533–539.

71. Hanrahan V, Currie MJ, Gunningham SP, Morrin HR, Scott PA,Robinson BA, et al. The angiogenic switch for vascularendothelial growth factor (VEGF)-A, VEGF-B, VEGF-C, andVEGF-D in the adenoma–carcinoma sequence during colorectalcancer progression. J Pathol 2003;200:183–194.

72. George ML, Tutton MG, Janssen F, Arnaout A, Abulafi AM,Eccles SA, et al. VEGF-A, VEGF-C, and VEGF-D in colorectalcancer progression. Neoplasia 2001;3:420–427.

73. Funaki H, Nishimura G, Harada S, Ninomiya I, Terada I,Fushida S, et al. Expression of vascular endothelial growth factorD is associated with lymph node metastasis in human colorectalcarcinoma. Oncology 2003;64:416–422.

74. White JD, Hewett PW, Kosuge D, McCulloch T, Enholm BC,Carmichael J, et al. Vascular endothelial growth factor-Dexpression is an independent prognostic marker for survival incolorectal carcinoma. Cancer Res 2002;62:1669–1675.

75. Elagoz S, Egilmez R, Koyuncu A, Muslehiddinoglu A, Arici S.The intratumoral microvessel density and expression of bFGFand nm23-H1 in colorectal cancer. Pathol Oncol Res 2006;12:21–27.

76. Chung YC, Hou YC, Chang CN, Hseu TH. Expression andprognostic significance of angiopoietin in colorectal carcinoma.J Surg Oncol 2006;94:631–638.

77. Schoppmann SF, Fenzl A, Nagy K, Unger S, Bayer G, Geleff S,et al. VEGF-C expressing tumor-associated macrophages inlymph node positive breast cancer: impact on lymphangiogenesisand survival. Surgery 2006;139:839–846.

78. Breslin JW, Gaudreault N, Watson KD, Reynoso R, Yuan SY,Wu MH. Vascular endothelial growth factor-C stimulates thelymphatic pump by a VEGF receptor-3-dependent mechanism.Am J Physiol Heart Circ Physiol 2007;293:H709–718.

79. Hoshida T, Isaka N, Hagendoorn J, di Tomaso E, Chen YL,Pytowski B, et al. Imaging steps of lymphatic metastasis revealsthat vascular endothelial growth factor-C increases metastasis byincreasing delivery of cancer cells to lymph nodes: therapeuticimplications. Cancer Res 2006;66:8065–8075.

80. Hirakawa S, Brown LF, Kodama S, Paavonen K, Alitalo K,Detmar M. VEGF-C-induced lymphangiogenesis in sentinellymph nodes promotes tumor metastasis to distant sites. Blood2007;109:1010–1017.

81. Qian CN, Berghuis B, Tsarfaty G, Bruch M, Kort EJ, Ditlev J,et al. Preparing the ‘soil’: the primary tumor induces vasculaturereorganization in the sentinel lymph node before the arrival ofmetastatic cancer cells. Cancer Res 2006;66:10365–10376.

82. Harrell MI, Iritani BM, Ruddell A. Tumor-induced sentinellymph node lymphangiogenesis and increased lymph flowprecede melanoma metastasis. Am J Pathol 2007;170:774–786.

83. Wiley HE, Gonzalez EB, Maki W, Wu MT, Hwang ST. Expres-sion of CC chemokine receptor-7 and regional lymph nodemetastasis of B16 murine melanoma. J Natl Cancer Inst2001;93:1638–1643.

84. Shields JD, Emmett MS, Dunn DB, Joory KD, Sage LM,Rigby H, et al. Chemokine-mediated migration of melanomacells towards lymphatics — a mechanism contributing tometastasis. Oncogene 2007;26:2997–3005.

85. Uchida D, Begum NM, Tomizuka Y, Bando T, Almofti A,Yoshida H, et al. Acquisition of lymph node, but not distantmetastatic potentials, by the overexpression of CXCR4 inhuman oral squamous cell carcinoma. Lab Invest 2004;84:1538–1546.

86. Muller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME,et al. Involvement of chemokine receptors in breast cancermetastasis. Nature 2001;410:50–56.

87. Kawada K, Hosogi H, Sonoshita M, Sakashita H, Man-abe T, Shimahara Y, et al. Chemokine receptor CXCR3 pro-motes colon cancer metastasis to lymph nodes. Oncogene2007;26:4679–4688.

88. Gunther K, Leier J, Henning G, Dimmler A, Weissbach R,Hohenberger W, et al. Prediction of lymph node metastasis incolorectal carcinoma by expression of chemokine receptor CCR7.Int J Cancer 2005;116:726–733.

89. St Croix B, Rago C, Velculescu V, Traverso G, Romans KE,Montgomery E, et al. Genes expressed in human tumorendothelium. Science 2000;289:1197–1202.

90. Nanda A, Buckhaults P, Seaman S, Agrawal N, Boutin P,Shankara S, et al. Identification of a binding partner for theendothelial cell surface proteins TEM7 and TEM7R. Cancer Res2004;64:8507–8511.

91. Nanda A, Carson-Walter EB, Seaman S, Barber TD, Stampfl J,Singh S, et al. TEM8 interacts with the cleaved C5 domain ofcollagen alpha 3(VI). Cancer Res 2004;64:817–820.

92. Zhang L, Giraudo E, Hoffman JA, Hanahan D, Ruoslahti E.Lymphatic zip codes in premalignant lesions and tumors. CancerRes 2006;66:5696–5706.

93. Clasper S, Royston D, Baban D, Cao Y, Ewers S, Butz S, et al.A novel gene expression profile in lymphatics associatedwith tumor growth and nodal metastasis. Cancer Res2008;68:7293–7303.

94. Yong C, Bridenbaugh EA, Zawieja DC, Swartz MA. Microarrayanalysis of VEGF-C responsive genes in human lymphaticendothelial cells. Lymphat Res Biol 2005;3:183–207.

95. Johnson LA, Clasper S, Holt AP, Lalor PF, Baban D, Jack-son DG. An inflammation-induced mechanism for leukocytetransmigration across lymphatic vessel endothelium. J Exp Med2006;203:2763–2777.

96. Wegmann F, Petri B, Khandoga AG, Moser C, Khandoga A,Volkery S, et al. ESAM supports neutrophil extravasation,activation of Rho, and VEGF-induced vascular permeability. JExp Med 2006;203:1671–1677.

97. Li DY, Sorensen LK, Brooke BS, Urness LD, Davis EC, Tay-lor DG, et al. Defective angiogenesis in mice lacking endoglin.Science 1999;284:1534–1537.

98. Fonsatti E, Maio M. Highlights on endoglin (CD105): from basicfindings towards clinical applications in human cancer. J TranslMed 2004;2:18.

99. Cao R, Brakenhielm E, Wahlestedt C, Thyberg J, Cao Y. Leptininduces vascular permeability and synergistically stimulatesangiogenesis with FGF-2 and VEGF. Proc Natl Acad Sci USA2001;98:6390–6395.

100. Joukov V, Pajusola K, Kaipainen A, Chilov D, Lahtinen I,Kukk E, et al. A novel vascular endothelial growth factor, VEGF-C, is a ligand for the Flt4 (VEGFR-3) and KDR (VEGFR-2)receptor tyrosine kinases. EMBO J 1996;15:290–298.

101. Valtola R, Salven P, Heikkila P, Taipale J, Joensuu H, Rehn M,et al. VEGFR-3 and its ligand VEGF-C are associated withangiogenesis in breast cancer. Am J Pathol 1999;154:1381–1390.

102. Partanen TA, Alitalo K, Miettinen M. Lack of lymphatic vascularspecificity of vascular endothelial growth factor receptor 3 in 185vascular tumors. Cancer 1999;86:2406–2412.

103. Oliver G, Sosa-Pineda B, Geisendorf S, Spana EP, Doe CQ,Gruss P. Prox 1, a prospero-related homeobox gene expressedduring mouse development. Mech Dev 1993;44:3–16.

104. Johnson NC, Dillard ME, Baluk P, McDonald DM, Harvey NL,Frase SL, et al. Lymphatic endothelial cell identity is reversibleand its maintenance requires Prox1 activity. Genes Dev2008;22:3282–3291.



105. Wigle JT, Harvey N, Detmar M, Lagutina I, Grosveld G,Gunn MD, et al. An essential role for Prox1 in the inductionof the lymphatic endothelial cell phenotype. EMBO J2002;21:1505–1513.

106. Breiteneder-Geleff S, Soleiman A, Horvat R, Amann G, Kowal-ski H, Kerjaschki D. [Podoplanin — a specific marker for lym-phatic endothelium expressed in angiosarcoma]. Verh Dtsch GesPathol 1999;83:270–275.

107. Nibbs RJ, Wylie SM, Yang J, Landau NR, Graham GJ. Cloningand characterization of a novel promiscuous human β-chemokinereceptor D6. J Biol Chem 1997;272:32078–32083.

108. Banerji S, Ni J, Wang SX, Clasper S, Su J, Tammi R,et al. LYVE-1, a new homologue of the CD44 glycoprotein,is a lymph-specific receptor for hyaluronan. J Cell Biol1999;144:789–801.


mechanisms of lymphatic metastasis in human colorectal adenocarcinoma

Documents