of January 20, 2012This information is current as

2004;173;7161-7169J Immunol and Peter PetzelbauerEmbacher, Manuela Pillinger, G. Scott Herron, Klaus Wolff Marion Gröger, Robert Loewe, Wolfgang Holnthoner, Robert CellsProx-1 and Podoplanin in Human Endothelial IL-3 Induces Expression of Lymphatic Markers

References

http://www.jimmunol.org/content/173/12/7161.full.html#related-urlsArticle cited in:  

http://www.jimmunol.org/content/173/12/7161.full.html#ref-list-1, 22 of which can be accessed free at:cites 52 articlesThis article

Subscriptions http://www.jimmunol.org/subscriptions

is online at The Journal of ImmunologyInformation about subscribing to

Permissions http://www.aai.org/ji/copyright.html

Submit copyright permission requests at

Email Alerts http://www.jimmunol.org/etoc/subscriptions.shtml/

Receive free email-alerts when new articles cite this article. Sign up at

Print ISSN: 0022-1767 Online ISSN: 1550-6606.Immunologists, Inc. All rights reserved.

by The American Association ofCopyright ©2004 9650 Rockville Pike, Bethesda, MD 20814-3994.The American Association of Immunologists, Inc.,

is published twice each month byThe Journal of Immunology

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

IL-3 Induces Expression of Lymphatic Markers Prox-1 andPodoplanin in Human Endothelial Cells1

Marion Groger,*† Robert Loewe,* Wolfgang Holnthoner,* Robert Embacher,*†

Manuela Pillinger,*† G. Scott Herron,‡ Klaus Wolff,*† and Peter Petzelbauer2*†

Factors determining lymphatic differentiation in the adult organism are not yet well characterized. We have made the observationthat mixed primary cultures of dermal blood endothelial cells (BEC) and lymphatic endothelial cells (LEC) grown under standardconditions change expression of markers during subculture: After passage 6, they uniformly express LEC-specific markers Prox-1and podoplanin. Using sorted cells, we show that LEC but not BEC constitutively express IL-3, which regulates Prox-1 andpodoplanin expression in LEC. The addition of IL-3 to the medium of BEC cultures induces Prox-1 and podoplanin. Blocking IL-3activity in LEC cultures results in a loss of Prox-1 and podoplanin expression. In conclusion, endogenous IL-3 is required tomaintain the LEC phenotype in culture, and the addition of IL-3 to BEC appears to induce transdifferentiation of BEC intoLEC. The Journal of Immunology, 2004, 173: 7161–7169.

I nterleukin-3 is a broadly acting regulatory protein of hemo-poietic cell differentiation. It stimulates proliferation, sur-vival, and differentiation of pluripotent hemopoietic stem

cells, and promotes expansion of cells that differentiate into maturecell lineages. Interestingly, IL-3 also has the capacity to divertdifferentiation of already committed cells. In the presence of IL-3,cells of the osteoclast lineage differentiate into cells of the mac-rophage lineage (1). IL-3 signals to cells via the IL-3R, which iscomposed of an IL-3-specific �-chain and a common �-chainshared with the GM-CSFR �-chain and the IL-5R �-chain (2, 3).

Endothelial cells (EC)3 express IL-3R �- and �-chains, and ex-pression can be increased by stimulation with TNF-� and/or IFN-�(4). EC respond to IL-3 by expressing IL-8 and E-selectin (5, 6).Moreover, IL-3 promotes endothelial tube formation and direc-tional migration (7). Cultured EC have been identified as a sourceof IL-3 (8). Whether EC-derived IL-3 has autocrine functions isunknown, as is the question of a role for IL-3 in EC differentiation.

EC invest two distinct vascular trees, one building the transportsystem for blood and the other for lymph. They differ by morphol-ogy and phenotype as reviewed recently by several authors (9–13).Established markers for human lymphatic cells are, for example,Prox-1 (14), podoplanin (15, 16) (also called T1� (17) or gp36(18)), vascular endothelial growth factor (VEGF)R-3 (19, 20), andlymphatic vessel endothelial hyaluronan receptor-1 (LYVE-1)(21). Also for blood EC (BEC), lineage-specific genes have beenidentified by transcriptional profiling (22–24), among constitu-

tively expressed markers are N-cadherin (22–24), pathologischeanatomie Leiden-endothelium Ag (PAL-E) (for microvessels) (16,25), or CXCR4 (22, 23). During embryogenesis, factors governinglymphatic differentiation are well described. The transcription fac-tor Prox-1 is the earliest detectable marker of lymphatic differen-tiation and is essential for differentiation of lymphatic EC (LEC)from the anterior cardinal vein endothelium (14, 26, 27). Lym-phangiogenesis is regulated by growth factors VEGF-C andVEGF-D and their receptor VEGFR-3 (20, 28–30). The trans-membrane mucoprotein podoplanin was shown to be important forthe correct lymphatic network formation (26). In the adult organ-ism, the situation is more complex and less defined. For example,in some tumors and in inflammation, the lymphatic growth factorreceptor VEGFR-3 is expressed on proliferating lymphatic andblood vessels, resulting in VEGF-C-induced lymph and blood ves-sel formation (31–34). This raises the question whether additionalfactors exist, which determine differentiation of LEC within theadult organism. In this study, we demonstrate that IL-3 regulatesexpression of LEC-specific markers Prox-1 and podoplanin.

Materials and MethodsCells

HUVEC were isolated as described (35) and cultured in IMDM with 20%FCS, glutamine (2 mM; Invitrogen Life Technologies, Carlsbad, CA), pen-icillin (100 U/ml), streptomycin (100 �g/ml; both Invitrogen Life Tech-nologies), and heparin/ECGS (PromoCell, Heidelberg, Germany).

Foreskin-derived microvascular EC were isolated and cultured as de-scribed (36). Briefly, foreskins were treated with dispase (Invitrogen LifeTechnologies) for 20 min at 37°C, followed by mechanically scraping ECwith a cell scraper. Cells were then seeded into fibronectin-coated wellsand cultured in EC growth medium (MV/ECGS; PromoCell). At passage 1,BEC and LEC were separated by magnetic sorting with an anti-podoplaninserum (15, 16).

Antibodies

First step Abs were as follows: anti-Prox-1 (RDI, Flanders, NJ), anti-macrophage mannose receptor (BD Pharmingen, San Jose, CA), PAL-E(RDI), anti-N-cadherin (BD Pharmingen), FITC-labeled CD31 (Immuno-tech, Marseille, France), anti-IL-3R �-chain Ab for FACS (clone 9F5), andblocking anti-IL-3R �-chain Ab (clone 7G3; both BD Pharmingen). As aspecificity control for the anti-podoplanin serum, we raised a mouse mono-clonal anti-human podoplanin Ab using peptide GASTGQPEDDTETTGLEGG (aa 22–40) as the Ag; a rabbit anti-human podoplanin serum was

*Department of Dermatology, Division of General Dermatology, Medical Universityof Vienna, and †Ludwig Boltzmann Institute for Angiogenesis, Microcirculation andInflammation, Vienna, Austria; and ‡Department Dermatology, Palo Alto MedicalClinic, Palo Alto, CA 94301

Received for publication May 26, 2004. Accepted for publication October 4, 2004.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.1 This work was supported by a research grant from the Stavros Niarchos Foundation.2 Address correspondence and reprint requests to Dr. Peter Petzelbauer, Depart-ment of Dermatology, Division of General Dermatology, Medical University ofVienna, Waehringer Guertel 18-20, A-1090 Vienna, Austria. E-mail address:Peter.Petzelbauer@meduniwien.ac.at3 Abbreviations used in this paper: EC, endothelial cell; BEC, blood EC; LEC, lymphaticEC; VEGF, vascular endothelial growth factor; LYVE, lymphatic vessel endothelialhyaluronan receptor; PAL-E, pathologische anatomie Leiden-endothelium Ag.

The Journal of Immunology

Copyright © 2004 by The American Association of Immunologists, Inc. 0022-1767/04/$02.00

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

a gift from Dr. D. Kerjaschki (Department of Pathology, Medical Univer-sity of Vienna). Isotype controls were purchased from Sigma-Aldrich (St.Louis, MO). Second step Abs were as follows: tetramethylrhodamine iso-thiocyanate-labeled goat anti-rabbit F(ab)2 (Jackson ImmunoResearchLaboratories, West Grove, PA). PE-labeled goat anti-rabbit F(ab)2 (BDPharmingen), Alexa 488-labeled anti-rabbit IgG (Molecular Probes, Lei-den, The Netherlands), FITC-labeled anti-mouse-IgG (Sigma-Aldrich), andHRP-labeled anti-mouse or anti-rabbit IgG (Bio-Rad, Hercules, CA).

FACS and laser scan analysis

For FACS, cells were detached from culture dishes using trypsin/EDTA.For laser scan imaging, cells were directly fixed on culture dishes withacetone/methanol (1:1) at �20°C for 10 min. Samples were then incubatedwith indicated first-step Abs followed by the appropriate fluorescence-la-beled second-step Abs. Isotype-matched control Abs were used in parallel.Bound fluorescence was analyzed by FACScan (BD Biosciences, Moun-tain View, CA) or laser scan microscope (LSM 520; Zeiss, Oberkochen,Germany).

RT-PCR

Cells were lysed in lysis buffer (10 mM NaOH-HEPES (pH 7.8), 1.5 mMMgCl, 10 mM KCl, 1 mM DTT, 0.1% Nonidet P-40, and 1 U of RNaseinhibitor (Roche, Basel, Switzerland)), and mRNA was isolated by usingoligo(dT) beads (Dynal, Oslo, Norway). RNA was reverse transcribed withSuperscript Reverse Transcriptase (Invitrogen Life Technologies) and hex-amer primers (Roche) for 90 min at 42°C. One microliter of cDNA wasthen used for PCR with primers for podoplanin (forward, 5�-GAA GGTGTC AGC TCT GCT CT-3�; reverse, 5�-ACG TTG GCA GGG CGTAA-3�; 35 cycles), Prox-1 (forward, 5�-AAG ACA GAG CCT CTC CTGAAT C-3�; reverse, 5�-TTG CAC TTC CCG AAT AAG GTG AT-3�; 30cycles), Lyve-1 (forward, 5�-GTG CTT CAG CCT GGT GTT G-3�; re-verse, 5�-GCT TGG ACT CTT GGA CTC TTC-3�; 30 cycles), flt-4 (for-ward, 5�-AGC TCT CAG AGC TCA GAA GAG-3�; reverse, 5�-TTC TCTCTC TCT GCT TCA GCT-3�; 30 cycles). Primers for GAPDH were pur-chased from BD Clontech (Palo Alto, CA). As control for genomic DNAcontamination, PCR was performed without reverse transcription. PCRconditions were 94°C for 30 s, 56°C for 40 s, and 72°C for 45 s. IL-3mRNA was analyzed by real-time PCR using primers from assay4327037T (Applied Biosystems, Foster City, CA) according to the instruc-tions of the manufacturer.

ELISA

Cell culture supernatants were collected, protease inhibitors added (PMSF,leupeptin, and aprotinin), and analyzed using an IL-3 ELISA kit from R&DSystems (Minneapolis, MN) according to the instructions of the manufac-turer. OD450 was measured.

Western blot

Western blotting was performed as described previously (37). Briefly, fol-lowing cell lysis with Tris lysis buffer (containing 10 mM Tris-HCl (pH7.5), 150 mM NaCl, 2 mM CaCl2, 1 mM PMSF, 10 �g/ml aprotinin, 15�g/ml leupeptin, 1% Nonidet P-40, and 1% Triton X-100), proteins wereseparated on SDS polyacrylamide gels and blotted on polyvinylidene di-fluoride membranes. Blots were then incubated with indicated first-stepAbs, followed by incubation with appropriate HRP-labeled second-stepAbs. Signals were visualized by chemiluminescence (Super ECL System;Amersham Biosciences, Buckinghamshire, U.K.) and documented on Hy-perfilm MP (Amersham Biosciences).

ResultsSkin-derived EC acquire a lymphatic phenotype duringsubculture

Normal human skin contains approximately equal numbers of po-doplanin-positive lymphatic vessels and podoplanin-negativeblood vessels (an example of an immunofluorescence staining ofhuman skin is shown in Fig. 1A). As a consequence, mechanicallyscraped EC consisted of LEC and BEC at a ratio of �1:1 (rangingfrom 2:1 to 1:2). A FACS analysis using anti-podoplanin Abs ofprimary isolated cells (passage 0 cells) is shown in Fig. 1B. Duringsubculture, the numbers of podoplanin-negative cells continuouslydecreased, and numbers of podoplanin-positive cells increased(Fig. 1B). Passage 6 cells consisted of 100% podoplanin-positive

cells. As a positive control, we used CD31 Abs, which reactedequally with LEC and BEC (Fig. 1B).

IL-3 is constitutively expressed in LEC

The molecular characterization of sorted LEC and BEC is shownin Fig. 1C. LEC are characterized by the expression of podoplanin,Flt-4, LYVE-1, Prox-1, and mannose receptor, and BEC by PAL-Eand N-cadherin expression. To determine factors inducing podo-planin expression of mixed (nonsorted) LEC/BEC in culture, wescreened cell culture supernatants of sorted LEC and BEC for cy-tokines. We found IL-3 consistently present in LEC and absent inBEC supernatants (Fig. 2A). HUVEC were analyzed for compar-ison and also found negative for IL-3 (see also Nilsen et al. (8)).Primary cultures, which, due to the isolation procedure, consistedof a mixture of LEC and BEC (mixed LEC/BEC), had slightlyhigher IL-3 release than in sorted LEC (by Student’s t test, thedifference compared with LEC was NS). IL-3 expression could beinduced following stimulation with TNF-�/IFN-� in LEC and inBEC, but IL-3 levels were �10 times higher in stimulated LECthan in stimulated BEC (Fig. 2A). To confirm results, IL-3 mRNAwas analyzed by real-time PCR: IL-3 mRNA was detectable inunstimulated LEC after a mean of 37 cycles, and was absent inunstimulated BEC even after 45 cycles (GAPDH was detectableafter 21 cycles). Fig. 2B gives an example of a representative semi-quantitative RT-PCR for IL-3.

It was previously shown that HUVEC express IL-3R �- and�-chain, and that expression is induced by TNF-�/IFN-� (4). Inthis study, we show that also LEC and BEC express IL-3R(slightly higher in LEC than in BEC), and that expression wasenhanced by TNF-�/IFN-�. Fig. 2C shows results for the �-chain;expression of the �-chain followed the pattern of the �-chain (datanot shown) (6, 38, 39). As a negative control, dermal fibroblastswere analyzed in parallel (Fig. 2C).

IL-3 up-regulates podoplanin and Prox-1 expression

Cell surface expression of podoplanin was analyzed by FACS.Following IL-3 stimulation, podoplanin expression was enhancedin LEC (3-fold). In BEC and HUVEC, IL-3 induced de novo ex-pression of podoplanin, albeit to a weaker extent (2-fold abovebaseline; Fig. 3). As a negative control, IL-3 was heat inactivated,which did not up-regulate podoplanin expression (Fig. 3). AlsoTNF-�/IFN-� induced podoplanin expression in LEC, BEC, andHUVEC. The addition of 1 �g/ml anti-IL-3R �-chain Ab (clone7G3, blocks signal transduction of the IL-3R (40)) inhibits TNF-�/IFN-�-induced podoplanin expression (Fig. 3), which indicatesthat a TNF-�/IFN-�-induced IL-3 release is responsible for theobserved podoplanin up-regulation. Following prestimulation withTNF-�/IFN-�, subsequent IL-3 stimulation further enhanced po-doplanin expression in LEC (4-fold). Also in BEC and HUVEC,TNF-�/IFN-� prestimulation followed by IL-3 was additive; po-doplanin expression was further induced (Fig. 3). Dermal fibro-blasts analyzed as a negative control neither expressed podoplaninat baseline nor following TNF-�/IFN-� and IL-3 stimulation(Fig. 3).

FACS data were confirmed by podoplanin Western blots usingwhole-cell lysates of LEC and BEC (Fig. 4A). Prox-1 expressionwas analyzed by immunofluorescence. In unstimulated LEC,Prox-1 was found in the nucleus in 100% of cells. In contrast, inunstimulated BEC, Prox-1 was undetectable, but was inducibleupon stimulation with IL-3 (Fig. 4B). Podoplanin and Prox-1mRNA expression was analyzed by semiquantitative RT-PCR andwere readily detectable in LEC at baseline and induced followingIL-3 stimulation (Fig. 4C). In contrast, unstimulated BEC andHUVEC were negative for podoplanin and Prox-1 mRNA, but was

7162 IL-3 INDUCES Prox-1 AND PODOPLANIN

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

FIGURE 1. A, Immunofluores-cence double-staining of normal hu-man skin samples with anti-podoplanin serum and CD31 Abs.Double-positive vessels representLEC and are marked by arrowheads.The two right panels show stainingwith the corresponding negative con-trols. B, FACS analysis for podopla-nin and CD31 expression of freshlyisolated EC cultures (p0), passage 1cells (p1), and passage 6 cells (p6)stained with anti-podoplanin serum(left panel), CD31 (middle panel),and controls (right panels). C,Marker profile of sorted BEC andLEC (passage 4 cells). FACS histo-grams are shown at the left side,Western blots at the right side, andRT-PCR at the bottom. Podoplaninwas analyzed with two Abs, an anti-podoplanin serum or a mouse mAb(marked by asterisks). Note that im-munoblots with anti-podoplanin se-rum reveal two bands of 38 and 28kDa (18), whereas the mouse mAbdetects a 38-kDa band only.

7163The Journal of Immunology

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

inducible following IL-3 stimulation (Fig. 4C). As a negative con-trol, dermal fibroblasts were analyzed and found negative for po-doplanin and Prox-1 mRNA (Fig. 4C).

Anti-IL-3R �-chain Abs inhibit Prox-1 and podoplaninexpression

To analyze the role of endogenous IL-3, sorted LEC were culturedin the presence of an anti-IL-3R �-chain Ab (1 �g/ml) (40). In thepresence of this blocking Ab, podoplanin expression in LEC de-creased (as analyzed by immunofluorescence LSM imaging; Fig.5A). After 24 h of treatment of LEC, islands of cells emerged,which had lost podoplanin expression, whereas other islandsshowed persisting podoplanin expression. Islands of podoplanin-positive cells became continuously smaller, and after 168 h, po-doplanin expression was almost completely lost in all LEC (Fig.5A). In the presence of a control Ab, cells remained positive for

podoplanin (Fig. 5A, first panel). Results were confirmed by FACSanalysis. Fig. 5B gives the cumulative result of three independentexperiments. After 168 h of anti-IL-3R� Ab treatment, �90% ofcells were negative for podoplanin expression, whereas an isotype-matched control Ab had no effect (Fig. 5B). After 168 h, the anti-IL-3R� Ab was removed by repeated washing and replaced byfresh medium. As shown in Fig. 5C, cells remained negative forpodoplanin even after withdrawal of the Ab. Results from immu-nofluorescence and FACS were confirmed by RT-PCR; podopla-nin and Prox-1 mRNA were undetectable following 72-h treatmentwith anti-IL-3R� Ab (Fig. 5B, inset).

Anti-IL-3R� Ab treatment also altered cell morphology of cul-tured LEC. A representative example of transmission light micros-copy images of LEC (cultured with isotype Ab), BEC (culturedwith isotype Ab), and LEC cultured with anti-IL-3R� Ab is shownin Fig. 5D. LEC appeared as flattened cells, whereas BEC had a

FIGURE 2. A, IL-3 secretion into the culturemedium of HUVEC, sorted BEC and LEC (pas-sage 4 cells), and primary cultures of freshly iso-lated skin-derived EC consisting of a mixture ofBEC and LEC (mixed BEC and LEC) as deter-mined by ELISA. Basal IL-3 release under stan-dard conditions is shown in the left panel. Theright panels show IL-3 release following stimu-lation with TNF-�/IFN-� (10 ng/ml each) for24 h (mean � SEM of three independent exper-iments performed each in triplicate; the differ-ence between BEC and LEC is significant, p �0.01; the difference between mixed cultures andLEC is NS). B, Semiquantitative RT-PCR forIL-3 mRNA expression in BEC and LEC isshown (35 cycles). As a control, GAPDHmRNA expression is shown (20 cycles). C,IL-3R �-chain expression as determined byFACS is shown. The left panels show IL-3R�-chain expression of indicated cells grown un-der standard conditions. The right panels showIL-3R �-chain expression following stimulationwith TNF-�/IFN-� (10 ng/ml each) for 12 h.Thin lines are isotype controls; bold lines arestaining with the anti-IL-3R� Ab.

7164 IL-3 INDUCES Prox-1 AND PODOPLANIN

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

more spindled morphology. In the presence of anti-IL-3R�, a pop-ulation of spindled cells emerged comparable to those seen in pureBEC cultures. These islands correspond to the podoplanin-nega-tive cells shown in Fig. 5A.

DiscussionProx-1 is the earliest lymphatic marker during embryonic devel-opment (14, 26) and is thought to be the critical gene programmingthe lymphatic phenotype (14, 23, 27, 41). Podoplanin expressionoccurs later and appears to be important for the lymphatic networkformation (23, 26, 41). Both molecules are constitutively ex-pressed on lymphatic endothelium in the adult organism, but fac-tors that induce or maintain their expression are currently un-known. In this study, we identified for the first time a solublefactor, IL-3, which is expressed in cultured LEC but not in BEC,and which is required to maintain expression of LEC-specific

markers Prox-1 and podoplanin. Blocking IL-3 effects leads to aloss of expression of these molecules in cultured LEC.

It should be noted that, during embryogenesis, IL-3 does notappear to be the critical factor for the formation of the lymphatictree, because IL-3-deficient animals have no severe lymphatic dys-function. However, current reports have not specifically searchedfor subtle changes in lymphatic function (42, 43); thus, the role ofIL-3 during development has to remain open.

Another issue concerns the question of why the expression ofIL-3 has not yet been identified as a LEC-specific gene in previousgene expression arrays (22–24). On the one hand, this might bedue to the reported instability of IL-3 mRNA, which complicatesdetection of IL-3 mRNA (44, 45). In contrast, this seeming dis-crepancy might be due to the very low IL-3 mRNA levels in un-stimulated LEC cultures, which may have escaped detection inarrays. However, constitutive IL-3 expression levels are generally

FIGURE 3. IL-3 induces podoplanin as determined by FACS analysis. Sorted LEC, sorted BEC, HUVEC (passage 4 cells), or dermal fibroblasts werecultured for 24 h in medium alone or in medium supplemented with IL-3 (25 ng/ml; 24 h), heat-inactivated IL-3, TNF-�/IFN-� (10 ng/ml each; 24 h),TNF-�/IFN-� plus anti-IL-3R� Ab 7G3 (1 �g/ml), or TNF-�/IFN-� (12 h) followed by IL-3 (24 h). Staining was performed with the control serum (rowat the top) or with anti-podoplanin serum (all others). Geometric mean fluorescence values are inserted into each panel and represent the mean of threeindependent experiments.

7165The Journal of Immunology

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

low in most cell types previously analyzed (46), and even such lowconcentrations are biologically active in settings of an autocrineIL-3 release. IL-3 supplemented to the medium requires muchhigher concentrations to achieve biological activity (46).

In our assays, committed cells were used, which were isolatedand sorted from human foreskins and thus already differentiatedinto either a BEC or a LEC phenotype. Even in such differentiatedcells, IL-3 was able to induce Prox-1 and podoplanin expression.Although IL-3-induced expression levels of Prox-1 and podopla-nin were much lower in BEC than in LEC, it involved the wholecell population, excluding growth selection as the basis of thisphenomenon. Moreover, phenotypic changes appeared within12–24 h, and population doubling times in EC are 36–48 h, whichalso argues against a growth selection phenomenon. We also showthat first-line inflammatory cytokines TNF-�/IFN-� induce podo-planin expression, and this also depended on IL-3 (because it isinhibited by anti-IL-3R� Abs). Moreover, TNF-�/IFN-� prestimu-

lation was additive to subsequent IL-3-induced Prox-1 and podo-planin expression. This additive effect is most likely based onTNF-�/IFN-�-induced IL-3R expression. Such a mechanism hasbeen previously described for IL-3-induced MHC class II expres-sion in HUVEC, where TNF-�/IFN-� prestimulation increasedIL-3R expression and subsequently MHC class II expression in-duced by IL-3 (4).

To date, IL-3 is the first soluble factor identified that is able toinduce Prox-1 and podoplanin expression in differentiated EC.This IL-3 effect was seen in LEC and in BEC, which raises theintriguing question whether IL-3-treated BEC transdifferentiateinto LEC. Based on current knowledge, the answer is complex,because (first) cultured LEC have a certain heterogeneous expres-sion of lymphatic markers (e.g., LYVE-1 or VEGFR-3 may beabsent). This phenomenon was reviewed by Saharinen et al. (47)and interpreted as the appearance of different LEC subpopulations.Second, the situation is more complicated in experiments where a

FIGURE 4. A, Western blotting for IL-3-inducedpodoplanin expression. IL-3 (25 ng/ml) was added tothe medium for 24 h. Cells were then lysed, and im-munoblots were performed with the polyclonal rabbitanti-podoplanin serum, which detects two differentgylcosylation forms, 28 and 38 kDa, respectively. As aloading control, immunoblots for CD31 expression areshown. B, Immunofluorescence laser scan images forProx-1 expression. Prox-1 is seen in a nuclear local-ization (red) in untreated LEC, and immunofluores-cence is enhanced following IL-3 stimulation. In con-trast, Prox-1 is absent in unstimulated BEC andinduced following IL-3 stimulation. To visualize cellmargins, cells were double-stained with CD31 Abs ingreen. C, Semiquantitative RT-PCR for podoplaninand Prox-1 expression in unstimulated cells or follow-ing IL-3-stimulation for indicated times is shown. Fi-broblasts were used as a negative control.

7166 IL-3 INDUCES Prox-1 AND PODOPLANIN

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

phenotypic switch from BEC into LEC was induced by, for ex-ample, overexpressing Prox-1 (23, 41) or by infection with Kar-posi sarcoma-associated herpes virus (HHV8) (48, 49). Prox-1overexpression in BEC induced only �20% of lymphatic genesand repressed only �40% of blood vessel genes. This is also truefor HHV8 infection of blood vessel endothelium, where some, butnot all, lymphatic markers were up-regulated, and some blood-specific markers disappeared. Surprisingly, even Prox-1 was not

induced in all experiments (48, 49). Thus the definition of a trueLEC or BEC phenotype still floats with accumulating data on thesecells (reviewed by Saharinen et al. (47)). In other words, we thinkthat the question whether IL-3 induces a true BEC-to-LEC con-version or a hybrid form cannot be answered at present.

With regard to in vivo effects of IL-3, others have shown newvessel formation in mice injected with IL-3-containing Matrigelplugs, but at the time of this report, phenotypic markers for LEC

FIGURE 5. Anti-IL-3R� Abs inhibit podoplanin and Prox-1 expression. A, LEC were cultured in the presence of anti-IL-3R� Ab 7G3 (1 �g/ml) forindicated times, and then double-stained with anti-podoplanin (red) and CD31 (green) Abs and analyzed by laser scan microscopy. The left image showspersisting podoplanin expression in the presence of a control Ab. After 24-h treatment with anti-IL-3R� Abs, islands of cells emerged, which had lostpodoplanin expression, whereas surrounding cells showed persisting podoplanin expression. After 168 h, podoplanin expression was almost completely lostin all LEC (right panel). B, FACS analysis for podoplanin expression in LEC treated with anti-IL-3R� or isotype control Abs. Geometric mean fluorescencelevels from three independent experiments are shown and expressed as the percentage of inhibition of podoplanin expression compared with untreated cells(mean � SD). The corresponding RT-PCR for podoplanin and Prox-1 mRNA expression is shown as an inset in B. C, Representative FACS histogram forpodoplanin expression in untreated LEC (first panel) and anti-IL-3R� Ab-treated LEC (168 h; second panel). Thereafter, the Ab was washed out, and cellswere cultured in medium alone for another 48 and 168 h, respectively. Podoplanin expression remained negative even after withdrawal of the Ab (thirdand fourth panels). D, Anti-IL-3R� Ab treatment changes cell morphology. Transmission microscopic images of LEC (first image) and BEC (third image)treated with a control Ab. The image in the middle shows LEC treated with 1 �g/ml anti-IL-3R� Ab for 48 h. The asterisk denotes flattened cells resemblingLEC, and the circles denote more spindled cells resembling BEC.

7167The Journal of Immunology

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

were yet not well characterized, and thus distinctions between LECand BEC were not made (7, 50). Moreover, this type of experimentwould not differentiate between lymphangiogenesis from pre-ex-isting lymph vessels and lymph vessel formation through transdif-ferentiation or a combination thereof. Anyway, the possibility forIL-3 to induce lymphangiogenesis raises the question whether in-creased IL-3 activity in tumors is associated with a high metastaticpotential. This was shown for rat mammary adenocarcinoma cellclones, which were highly metastatic when they produced highlevels of IL-3, but were poorly metastatic when they had no IL-3activity (51). As a natural source of IL-3, tumor-infiltrating lym-phocytes have to be considered (50, 52). Finally, the effects of theproinflammatory cytokines TNF-� and IFN-� on Prox-1 and po-doplanin expression raises the possibility that lymphangiogenesisobserved in situations of chronic inflammation (53, 54) is in partmediated through transdifferentiation of BEC into LEC.

In conclusion, we show that endogenous IL-3 is required tomaintain the LEC phenotype in culture, and the addition of IL-3 toBEC induces expression of LEC-specific markers. Our data sup-port the possibility that BECs can reprogram their phenotype intoLECs, and IL-3 may contribute to lymphangiogenesis in tumors orin settings of chronic inflammation.

AcknowledgmentsWe thank Dr. D. Kerjaschki (Department of Pathology, Medical Universityof Vienna) for critical discussions and for the anti-podoplanin serum.

References1. Khapli, S. M., L. S. Mangashetti, S. D. Yogesha, and M. R. Wani. 2003. IL-3 acts

directly on osteoclast precursors and irreversibly inhibits receptor activator ofNF-�B ligand-induced osteoclast differentiation by diverting the cells to macro-phage lineage. J. Immunol. 171:142.

2. Geijsen, N., L. Koenderman, and P. J. Coffer. 2001. Specificity in cytokine signaltransduction: lessons learned from the IL-3/IL-5/GM-CSF receptor family. Cy-tokine Growth Factor Rev. 12:19.

3. Martinez-Moczygemba, M., and D. P. Huston. 2003. Biology of common � re-ceptor-signaling cytokines: IL-3, IL-5, and GM-CSF. J. Allergy Clin. Immunol.112:653.

4. Korpelainen, E. I., J. R. Gamble, W. B. Smith, M. Dottore, M. A. Vadas, andA. F. Lopez. 1995. Interferon-� upregulates interleukin-3 (IL-3) receptor expres-sion in human endothelial cells and synergizes with IL-3 in stimulating majorhistocompatibility complex class II expression and cytokine production. Blood86:176.

5. Brizzi, M. F., G. Garbarino, P. R. Rossi, G. L. Pagliardi, C. Arduino,G. C. Avanzi, and L. Pegoraro. 1993. Interleukin 3 stimulates proliferation andtriggers endothelial-leukocyte adhesion molecule 1 gene activation of humanendothelial cells. J. Clin. Invest. 91:2887.

6. Korpelainen, E. I., J. R. Gamble, W. B. Smith, G. J. Goodall, S. Qiyu,J. M. Woodcock, M. Dottore, M. A. Vadas, and A. F. Lopez. 1993. The receptorfor interleukin 3 is selectively induced in human endothelial cells by tumor ne-crosis factor � and potentiates interleukin 8 secretion and neutrophil transmigra-tion. Proc. Natl. Acad. Sci. USA 90:11137.

7. Dentelli, P., L. Del Sorbo, A. Rosso, A. Molinar, G. Garbarino, G. Camussi,L. Pegoraro, and M. F. Brizzi. 1999. Human IL-3 stimulates endothelial cellmotility and promotes in vivo new vessel formation. J. Immunol. 163:2151.

8. Nilsen, E. M., F. E. Johansen, F. L. Jahnsen, K. E. Lundin, T. Scholz,P. Brandtzaeg, and G. Haraldsen. 1998. Cytokine profiles of cultured microvas-cular endothelial cells from the human intestine. Gut 42:635.

9. Alitalo, K., and P. Carmeliet. 2002. Molecular mechanisms of lymphangiogenesisin health and disease. Cancer Cell 1:219.

10. Baldwin, M. E., S. A. Stacker, and M. G. Achen. 2002. Molecular control oflymphangiogenesis. BioEssays 24:1030.

11. Jackson, D. G. 2003. The lymphatics revisited: new perspectives from the hya-luronan receptor LYVE-1. Trends Cardiovasc. Med. 13:1.

12. Oliver, G., and M. Detmar. 2002. The rediscovery of the lymphatic system: oldand new insights into the development and biological function of the lymphaticvasculature. Genes Dev. 16:773.

13. Pepper, M. S., and M. Skobe. 2003. Lymphatic endothelium: morphological,molecular and functional properties. J. Cell Biol. 163:209.

14. Wigle, J. T., N. Harvey, M. Detmar, I. Lagutina, G. Grosveld, M. D. Gunn,D. G. Jackson, and G. Oliver. 2002. An essential role for Prox1 in the inductionof the lymphatic endothelial cell phenotype. EMBO J. 21:1505.

15. Breiteneder-Geleff, S., A. Soleiman, H. Kowalski, R. Horvat, G. Amann,E. Kriehuber, K. Diem, W. Weninger, E. Tschachler, K. Alitalo, andD. Kerjaschki. 1999. Angiosarcomas express mixed endothelial phenotypes ofblood and lymphatic capillaries: podoplanin as a specific marker for lymphaticendothelium. Am. J. Pathol. 154:385.

16. Kriehuber, E., S. Breiteneder-Geleff, M. Groeger, A. Soleiman,S. F. Schoppmann, G. Stingl, D. Kerjaschki, and D. Maurer. 2001. Isolation andcharacterization of dermal lymphatic and blood endothelial cells reveal stable andfunctionally specialized cell lineages. J. Exp. Med. 194:797.

17. Ma, T., B. Yang, M. A. Matthay, and A. S. Verkman. 1998. Evidence against arole of mouse, rat, and two cloned human T1� isoforms as a water channel or aregulator of aquaporin-type water channels. Am. J. Respir. Cell Mol. Biol.19:143.

18. Zimmer, G., F. Oeffner, V. Von Messling, T. Tschernig, H. J. Groness,H. D. Klenk, and G. Herrler. 1999. Cloning and characterization of gp36, ahuman mucin-type glycoprotein preferentially expressed in vascular endothe-lium. Biochem. J. 341:277.

19. Kaipainen, A., J. Korhonen, T. Mustonen, V. W. van Hinsbergh, G. H. Fang,D. Dumont, M. Breitman, and K. Alitalo. 1995. Expression of the fms-like ty-rosine kinase 4 gene becomes restricted to lymphatic endothelium during devel-opment. Proc. Natl. Acad. Sci. USA 92:3566.

20. Karkkainen, M. J., T. Makinen, and K. Alitalo. 2002. Lymphatic endothelium: anew frontier of metastasis research. Nat. Cell Biol. 4:E2.

21. Banerji, S., J. Ni, S. X. Wang, S. Clasper, J. Su, R. Tammi, M. Jones, andD. G. Jackson. 1999. LYVE-1, a new homologue of the CD44 glycoprotein, is alymph-specific receptor for hyaluronan. J. Cell Biol. 144:789.

22. Hirakawa, S., Y. K. Hong, N. Harvey, V. Schacht, K. Matsuda, T. Libermann,and M. Detmar. 2003. Identification of vascular lineage-specific genes by tran-scriptional profiling of isolated blood vascular and lymphatic endothelial cells.Am. J. Pathol. 162:575.

23. Petrova, T. V., T. Makinen, T. P. Makela, J. Saarela, I. Virtanen, R. E. Ferrell,D. N. Finegold, D. Kerjaschki, S. Yla-Herttuala, and K. Alitalo. 2002. Lymphaticendothelial reprogramming of vascular endothelial cells by the Prox-1 homeoboxtranscription factor. EMBO J. 21:4593.

24. Podgrabinska, S., P. Braun, P. Velasco, B. Kloos, M. S. Pepper, D. G. Jackson,and M. Skobe. 2002. Molecular characterization of lymphatic endothelial cells.Proc. Natl. Acad. Sci. USA 99:16069.

25. Schlingemann, R. O., G. M. Dingjan, J. J. Emeis, J. Blok, S. O. Warnaar, andD. J. Ruiter. 1985. Monoclonal antibody PAL-E specific for endothelium. Lab.Invest. 52:71.

26. Schacht, V., M. I. Ramirez, Y. K. Hong, S. Hirakawa, D. Feng, N. Harvey,M. Williams, A. M. Dvorak, H. F. Dvorak, G. Oliver, and M. Detmar. 2003.T1�/podoplanin deficiency disrupts normal lymphatic vasculature formation andcauses lymphedema. EMBO J. 22:3546.

27. Wilting, J., M. Papoutsi, B. Christ, K. H. Nicolaides, C. S. von Kaisenberg,J. Borges, G. B. Stark, K. Alitalo, S. I. Tomarev, C. Niemeyer, and J. Rossler.2002. The transcription factor Prox1 is a marker for lymphatic endothelial cellsin normal and diseased human tissues. FASEB J. 16:1271.

28. Karkkainen, M. J., P. Haiko, K. Sainio, J. Partanen, J. Taipale, T. V. Petrova,M. Jeltsch, D. G. Jackson, M. Talikka, H. Rauvala, et al. 2004. Vascular endo-thelial growth factor C is required for sprouting of the first lymphatic vessels fromembryonic veins. Nat. Immunol. 5:74.

29. Makinen, T., L. Jussila, T. Veikkola, T. Karpanen, M. I. Kettunen,K. J. Pulkkanen, R. Kauppinen, D. G. Jackson, H. Kubo, S. Nishikawa, et al.2001. Inhibition of lymphangiogenesis with resulting lymphedema in transgenicmice expressing soluble VEGF receptor-3. Nat. Med. 7:199.

30. Veikkola, T., L. Jussila, T. Makinen, T. Karpanen, M. Jeltsch, T. V. Petrova,H. Kubo, G. Thurston, D. M. McDonald, M. G. Achen, et al. 2001. Signalling viavascular endothelial growth factor receptor-3 is sufficient for lymphangiogenesisin transgenic mice. EMBO J. 20:1223.

31. Kubo, H., T. Fujiwara, L. Jussila, H. Hashi, M. Ogawa, K. Shimizu, M. Awane,Y. Sakai, A. Takabayashi, K. Alitalo, et al. 2000. Involvement of vascular en-dothelial growth factor receptor-3 in maintenance of integrity of endothelial celllining during tumor angiogenesis. Blood 96:546.

32. Paavonen, K., P. Puolakkainen, L. Jussila, T. Jahkola, and K. Alitalo. 2000.Vascular endothelial growth factor receptor-3 in lymphangiogenesis in woundhealing. Am. J. Pathol. 156:1499.

33. Partanen, T. A., K. Alitalo, and M. Miettinen. 1999. Lack of lymphatic vascularspecificity of vascular endothelial growth factor receptor 3 in 185 vascular tu-mors. Cancer 86:2406.

34. Partanen, T. A., J. Arola, A. Saaristo, L. Jussila, A. Ora, M. Miettinen,S. A. Stacker, M. G. Achen, and K. Alitalo. 2000. VEGF-C and VEGF-D ex-pression in neuroendocrine cells and their receptor, VEGFR-3, in fenestratedblood vessels in human tissues. FASEB J. 14:2087.

35. Petzelbauer, P., C. A. Watson, S. E. Pfau, and J. S. Pober. 1995. IL-8 and an-giogenesis: evidence that human endothelial cells lack receptors and do not re-spond to IL-8 in vitro. Cytokine 7:267.

36. Petzelbauer, P., J. R. Bender, J. Wilson, and J. S. Pober. 1993. Heterogeneity ofdermal microvascular endothelial cell antigen expression and cytokine respon-siveness in situ and in cell culture. J. Immunol. 151:5062.

37. Matsumura, T., K. Wolff, and P. Petzelbauer. 1997. Endothelial cell tube forma-tion depends on cadherin 5 and CD31 interactions with filamentous actin. J. Im-munol. 158:3408.

38. Colotta, F., F. Bussolino, N. Polentarutti, A. Guglielmetti, M. Sironi,E. Bocchietto, M. De Rossi, and A. Mantovani. 1993. Differential expression ofthe common � and specific � chains of the receptors for GM-CSF, IL-3, and IL-5in endothelial cells. Exp. Cell Res. 206:311.

39. Korpelainen, E. I., J. R. Gamble, M. A. Vadas, and A. F. Lopez. 1996. IL-3receptor expression, regulation and function in cells of the vasculature. Immunol.Cell Biol. 74:1.

40. Sun, Q., J. M. Woodcock, A. Rapoport, F. C. Stomski, E. I. Korpelainen,C. J. Bagley, G. J. Goodall, W. B. Smith, J. R. Gamble, M. A. Vadas, and

7168 IL-3 INDUCES Prox-1 AND PODOPLANIN

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from

A. F. Lopez. 1996. Monoclonal antibody 7G3 recognizes the N-terminal domainof the human interleukin-3 (IL-3) receptor �-chain and functions as a specificIL-3 receptor antagonist. Blood 87:83.

41. Hong, Y. K., N. Harvey, Y. H. Noh, V. Schacht, S. Hirakawa, M. Detmar, andG. Oliver. 2002. Prox1 is a master control gene in the program specifying lym-phatic endothelial cell fate. Dev. Dyn. 225:351.

42. Lantz, C. S., J. Boesiger, C. H. Song, N. Mach, T. Kobayashi, R. C. Mulligan,Y. Nawa, G. Dranoff, and S. J. Galli. 1998. Role for interleukin-3 in mast-cell andbasophil development and in immunity to parasites. Nature 392:90.

43. Guthridge, M. A., F. C. Stomski, D. Thomas, J. M. Woodcock, C. J. Bagley,M. C. Berndt, and A. F. Lopez. 1998. Mechanism of activation of the GM-CSF,IL-3, and IL-5 family of receptors. Stem Cells 16:301.

44. Hirsch, H. H., A. P. Nair, V. Backenstoss, and C. Moroni. 1995. Interleukin-3mRNA stabilization by a trans-acting mechanism in autocrine tumors lackinginterleukin-3 gene rearrangements. J. Biol. Chem. 270:20629.

45. Stoecklin, G., X. F. Ming, R. Looser, and C. Moroni. 2000. Somatic mRNAturnover mutants implicate tristetraprolin in the interleukin-3 mRNA degradationpathway. Mol. Cell. Biol. 20:3753.

46. Jiang, X., A. Lopez, T. Holyoake, A. Eaves, and C. Eaves. 1999. Autocrineproduction and action of IL-3 and granulocyte colony-stimulating factor inchronic myeloid leukemia. Proc. Natl. Acad. Sci. USA 96:12804.

47. Saharinen, P., T. Tammela, M. Karkkainen, and K. Alitalo. 2004. Lymphaticvasculature: development, molecular regulation and role in tumor metastasis andinflammation. Trends Immunol. 25:387.

48. Hong, Y., K. Foreman, J. W. Shin, S. Hirakawa, C. L. Curry, D. R. Sage,T. Libermann, B. J. Dezube, J. D. Fingeroth, and M. Detmar. 2004. Lymphatic

reprogramming of blood vascular endothelium by Kaposi sarcoma-associatedherpesvirus. Nat. Genet. 36:683.

49. Wang, H., M. W. B. Trotter, D. Lagos, D. Bourboulia, S. Henderson, T. Makinen,S. Elliman, A. M. Flanagan, K. Alitalo, and C. Boshoff. 2004. Kaposi sarcomaherpesvirus-induced cellular reprogramming contributes to the lymphatic endo-thelial gene expression in Kaposi sarcoma. Nat. Genet. 36:687.

50. Dentelli, P., A. Rosso, C. Calvi, B. Ghiringhello, G. Garbarino, G. Camussi,L. Pegoraro, and M. F. Brizzi. 2004. IL-3 affects endothelial cell-mediatedsmooth muscle cell recruitment by increasing TGF� activity: potential role intumor vessel stabilization. Oncogene 23:1681.

51. McGary, C. T., M. E. Miele, and D. R. Welch. 1995. Highly metastatic 13762NFrat mammary adenocarcinoma cell clones stimulate bone marrow by secretion ofgranulocyte-macrophage colony-stimulating factor/interleukin-3 activity.Am. J. Pathol. 147:1668.

52. Kharkevitch, D. D., D. Seito, G. C. Balch, T. Maeda, C. M. Balch, and K. Itoh.1994. Characterization of autologous tumor-specific T-helper 2 cells in tumor-infiltrating lymphocytes from a patient with metastatic melanoma. Int. J. Cancer58:317.

53. Geleff, S., S. F. Schoppmann, and G. Oberhuber. 2003. Increase in podoplanin-expressing intestinal lymphatic vessels in inflammatory bowel disease. VirchowsArch. 442:231.

54. Kerjaschki, D., H. M. Regele, I. Moosberger, K. Nagy-Bojarski, B. Watschinger,A. Soleiman, P. Birner, S. Krieger, A. Hovorka, G. Silberhumer, et al. 2004.Lymphatic neoangiogenesis in human kidney transplants is associated with im-munologically active lymphocytic infiltrates. J. Am. Soc. Nephrol. 15:603.

7169The Journal of Immunology

on January 20, 2012w

ww

.jimm

unol.orgD

ownloaded from