Soluble Carcinoembryonic Antigen Activates Endothelial Cells and Tumor Angiogenesis

Home , Carcinoembryonic antigen

Author Manuscript Published OnlineFirst on October 11, 2013; DOI: 10.1158/0008-5472.CAN-13-0123 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Soluble carcinoembryonic antigen activates endothelial cells and tumor angiogenesis

Kira H. Bramswig,1,5 Marina Poettler,1,5 Matthias Unseld,1 Friedrich Wrba,2 Pavel Uhrin,3 Wolfgang Zimmermann,4 Christoph C. Zielinski,1 Gerald W. Prager1,*

1 Clinical Division of Oncology, Department of Medicine I and Cancer Center, Medical University of Vienna, Austria

2 Institute of Clinical Pathology, Medical University of Vienna, Austria

3 Department of Vascular Biology and Thrombosis Research, Centre for Bio-Molecular Medicine and Pharmacology, Medical University of Vienna, Austria

4 Tumor Immunology Laboratory, LIFE-Center, Klinikum Grosshadern, Ludwig-Maximilians- University Munich

5 These authors contributed equally to this work

*Correspondence:

Dr. Gerald Prager Waehringer Guertel 18-20, A-1090 Vienna, Austria Tel: 0043-1-40400-4450 Fax: 0043-1-40400-4451 [email protected]

Running Title: CEA affects angiogenesis

Conflicts of interest: None

Abstract: 183 words

Text-Body without References: 4998 words

Keywords: carcinoembryonic antigen, endothelial cells, integrins, tumor-angiogenesis

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2013 American Association for Cancer Research. Author Manuscript Published OnlineFirst on October 11, 2013; DOI: 10.1158/0008-5472.CAN-13-0123 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Abstract

Carcinoembryonic antigen (CEA, CD66e, CEACAM-5) is a cell-surface bound

glycoprotein overexpressed and released by many solid tumors that has an autocrine

function in cancer cell survival and differentiation. Soluble CEA released by tumors is

present in the circulation of cancer patients, where it is used as a marker for cancer

progression, but whether this form of CEA exerts any effects in the tumor

microenvironment is unknown. Here we present evidence that soluble CEA is

sufficient to induce pro-angiogenic endothelial cell behaviors including adhesion,

spreading, proliferation and migration in vitro and tumor microvascularisation in vivo.

CEA-induced activation of endothelial cells was dependent on integrin beta-3 signals

that activate the FAK and c-Src kinase and their downstream MEK/ERK and PI3K/Akt

effector pathways. Notably, while interference with VEGF signaling had no effect on

CEA-induced endothelial cell activation, downregulation with the CEA receptor in

endothelial cells attenuated CEA-induced signaling and tumor angiogenesis.

Corroborating these results clinically, we found that tumor microvascularization was

higher in colorectal cancer patients exhibiting higher serum levels of soluble CEA.

Together, our results elucidate a novel function for soluble CEA in tumor

angiogenesis.

Introduction

Tumor angiogenesis is induced when the net balance of pro- and anti-angiogenic

molecules is tipped in favor of angiogenesis, resulting in the ‘angiogenic switch' (1,2).

Targeting angiogenesis have led to improved prognosis in many cancers including

those originating from colon, lung, brain, ovary and kidney (3). As vascular

endothelial growth factor (VEGF) represent the main pro-angiogenic stimulus (1), this

system is currently in focus for therapeutic interventions and has led to several FDA-

approved drugs in the treatment for cancer (4). Nevertheless, the use of specific

VEGF-targeting drugs has been shown to be effective only for certain patients and for

a limited duration of time (5,6), which might lie in the fact that tumor angiogenesis is

not only induced by VEGF, but also by a variety of other factors (3),(7,8).

In order to monitor adenocarcinoma growth as well as efficacy of its treatment,

soluble CD66e/CEA - the product of the CEACAM5-gene (9) - is in routine clinical

use. Under physiologic conditions only low amounts of soluble CEA (≤5 ng/ml) can

be detected in serum. However, in many different cancers CEA is highly upregulated,

but no clear biological role has emerged so far. Only recently, autocrine CEA

signaling in cancer cells was described to inhibit tumor cell differentiation and

apoptosis in vitro(10,11-12) and in vivo (13). Although CEA is highly released by

tumor cells, paracrine effects of soluble CEA have hitherto not been described. As

CEA is a GPI-anchored protein lacking a cytoplasmic domain, transmembrane

interaction partners are required for mediating intracellular signal transduction. In this

context, a co-localization of CEA with integrins in lipid rafts was observed, which

modulated integrin-dependent signaling pathways, including ILK, PI3kinase, and AKT

(14,15). 3

We were referred to a potential paracrine function of CEA in a previous CEA-

mimotope immunization study (16). In this particular study, CEA was the immunologic

target of a CEA-directed mimotope vaccination resulting in a significantly reduced

growth of CEA-expressing tumor transplants (Meth-A/CEA). We now observed that

CEA-mimotope immunized mice had significantly lower vascular density as well as

significantly decreased vessel area in tumor transplants, when compared to tumors in

non-CEA-vaccinated mice (Supplemental Fig.1). These data suggested CEA as a

potential novel stimulus in tumor-angiogenesis.

In the present study, we aimed to analyze whether CEA exerts a so far unknown

function in tumor-angiogenesis. In detail, we aimed 1.) to analyze any direct effects of

exogenous CEA on endothelial cells by in vivo angiogenesis assays as well as by in

vitro analyzes of endothelial cell behavior, 2.) to characterize potential CEA-induced

effects on intracellular signal transduction in endothelial cells, 3.) to identify

mechanisms responsible for CEA-induced endothelial cell activation and, finally, 4.)

to analyze the role of CEA-expression in gain and loss of function in in vivo tumor

transplant models.

Materials and Methods

A detailed description is given in the Supporting Information.

Immunoblotting and Western blots

Standard procedures were used(17,18).

Integrin Activity Assays

Binding of 2mM FITC-labeled cyclic-RGD (Anaspect, CA, USA) (Supplemental Fig.4)

or WOW-1 Fab(19) (100nM) (kindly gift from Sanford Shattil, UCSD) to human

microvascular endothelial cells (HUMECs) was assessed by flow cytometry

(FACScan, BD). Briefly, HUMECs (serum-free condition, 1%BSA) were incubated

with CEA (10ng/ml, 30min) or MnCl2 (Mn++,1mM) to activate integrins(20). Alexa488-

conjugated anti-mouse IgG (1:500) as secondary antibody; propidium iodide (2μg/ml,

Sigma-Aldrich, Austria). Three independent experiments were performed.

Immunohistochemistry

Immunohistochemistry was performed as described previously(21). Endothelial cell

staining was performed by anti-endomucin (eBioscience, CA)(displayed in figures),

anti-CD31 (Santa Cruz-Biotechnology, CA) or anti-vWF (Abcam, UK) antibodies.

CRC tumor-tissue sections were derived from the Department of Clinical Pathology,

MUW. As expected, no soluble CEA was detected on endothelial cells in tumor-tissue

sections as standard procedures of immunohistochemical-analyses include acid

treatment.

In vivo tumor-transplant models

Cell lines used were obtained from ATCC (Manassas, VA). Meth-A/CEA cells or

control Meth-A/wt cells were transplanted into BALB/c-mice (Institute of Laboratory-

Animal-Science, Austria) by subcutaneous injection into the right flank, 6

animals/group. Alternatively, NSG-mice (NOD-scid-gamma(NOD.Cg-

PrkdcscidIl2rgtm1Wjl/SzJ)) from Jackson-Laboratory (Maine, USA) were used to exclude

NK-cell or cytokines involvement for CEA effects(22). In xenotransplant-assays,

Caki2/mock(control) or Caki2/CEA(stably-overexpressing CEA) tumor cells were

injected into nude-mice (Institute of Laboratory-Animal-Science, Austria). After 8 or

21 days mice were sacrificed and tumors were explanted. Paraffin-embedded tumor-

tissue samples from the spontaneous-gastric-tumor model were derived from

CEA424/SV40T-antigen transgenic mice ± a human CEA-transgene(23-25).

Carcinomas were obtained from 8 different 13 week-old mice (4 CEA424/Tag- and 4

CEA424/Tag-CEA-double transgenic mice). Only tumors of same sizes were

analyzed for angiogenesis (as described above).

Aortic ring assay

Aortic ring explants of 6-12 weeks old C57BL/6-mice were placed on a growth factor-

reduced matrigel supplemented with 3%BSA containing medium ± CEA (100ng/ml)

or VEGF (30ng/ml). After six days a software-based (ImageJ 1.32) computer-assisted

method, which measured the number and length of vessels, was used. Supernatants

of MethA/CEA cells adjusted to 100ng/ml CEA were incubated with either IgG- or

anti-CEA-beads before used in the aortic-ring assay. Similar method as used in

Suppl.Fig.3C; CEA-level of MethA/wt: 0 ng/ml. Quantification is given in bars; error

bars: SD, n=3

Directed in vivo Angiogenesis-Assay (DIVAA™)

DIVAA (Trevigen, MD) was performed according to the manufacturer’s protocol with

modifications: CEA(1µg/ml) or BSA 1%(control) was added. Two reactors were

implanted into the flank of 6-8 weeks old, anaesthetized (Ketamin,Xylen) BL6-mice, 8

reactors/group. On day eleven mice were sacrificed and analyzed by

immunofluorescence, 7µm cryostat-sections. Functional (perfused) vessels were

identified by Hoechst 33342-dye staining (Supporting Information) and analyzed by

Olympus-AX70 microscope and quantified by Cell®-imaging software (Olympus-Soft-

Imaging Solutions).

Proliferation and Endothelial Cell Trans-Migration Assay

Serum deprived HUMECs were incubated with VEGF165(10ng/ml) or CEA(conc. as

indicated) ± inhibitors. The number of attached endothelial cells per unit was counted

after 24hours. Alternatively, MTT-assays (Promega, WI) were performed according to

manufactures instruction (n≥3). Cell migration was assayed in a modified Boyden-

chamber system by using transwell-membranes (8 µm) coated with 1%gelatin

(Supporting Information).

Cell Adhesion and Cell Spreading Assays

Assays of cell spreading on 20µg/ml fibrinogen (Fg) or 10µg/ml fibronectin (Fb)

(Sigma-Aldrich, Austria) were performed as described previously(26).

Statistical Analysis

Statistical significance was analyzed by un-/paired t-Test when one group was

compared with the control group. To compare two or more groups with the control

group “one way analysis of variance” and Dunnett’s tests as post tests were used.

Significance was assessed to p-values of less 0.05.

Results

CEA directly affects pro-angiogenic endothelial cell behavior

Based on the observation of reduced angiogenesis in CEA-immunized mice(16) and

the clinical correlation of high CEA with worse prognosis in cancer patients(27), we

postulated that tumor-associated CEA might affect the tumor-microenvironment,

thereby activating endothelial cells and enhance angiogenesis. To test this

possibility, we first analyzed potential effects of CEA on endothelial cell proliferation,

an essential step in angiogenesis. We found that CEA induced a significant dose-

dependent increase in endothelial cell proliferation, starting at concentrations of ≥ 80

ng / ml (Fig. 1A). Notably, simultaneous stimulation of endothelial cells with CEA and

VEGF led to a partially additive effect on proliferative activity (Fig.1B). As endothelial

cell migration represents another important step in angiogenesis, we next analyzed

direct effects of CEA on the migratory response using an in vitro transwell-migration

assay. CEA led to an increase in endothelial cell trans-migration starting at

concentrations of 10 ng/ml (Fig.1C). Concomitant stimulation by CEA and VEGF

resulted in an additional migratory response (not shown). Based upon results of the

in vitro endothelial cell transmigration assay, we studied endothelial cell invasion in

vivo using the in vivo Matrigel™-plug-assay (Supplementary Fig.2A) showing a

2.36±0.29-fold or 2.90±0.59-fold increase in invaded endothelial cells whenever CEA

(1 ug/ml) or VEGF (300 ng/ml) was embedded when compared to control (1%BSA)

(p<0.01). In an aortic ring ex vivo assay, we found that the addition of CEA or VEGF

led to increase in the mean length of sprouts (128.0+/-18.1µm or 155.7+/-23.1µm)

when compared to 3%BSA (control). Addition of supernatants from CEA-expressing

tumor cells (MethA/CEA) provoked endothelial sprouting, which was reverted by a 9

pulldown of CEA in supernatants (Fig.1D). We next introduced a directed in vivo

angiogenesis assay (Fig.1E). Incorporation of CEA (1 ug/ml) into the matrix led to an

increase in invaded endothelial cells after eleven days. Furthermore, the number of

Hoechst 33342 positive cells - representing the area of perfusion - was significantly

higher whenever CEA was present (CEA: 0.768±0.236 mm2 vs. control: 0.120±0.012

mm2 for control, p=0.02). Taken together, these data suggest that CEA directly

induces pro-angiogenic endothelial cell behavior in vitro and in vivo.

CEA induces intracellular signal transduction in endothelial cells

Next, we determined whether CEA mediates its effects on endothelial cells by

activating intracellular signaling pathways. First, we tested its effects on the mitogen-

activated protein kinase (MAPK) / extracellular signal-regulated kinase (ERK,

p44/p42 MAPK) signaling-pathway, which mainly regulates cellular proliferation and

differentiation(28). We found that CEA induced p44/p42 MAPK phosphorylation in a

mitogen extracellular kinase-1/2 (MEK1/2)-dependent manner, as phosphorylation

was blocked by the MEK-inhibitor U0126 (Fig.2A). Thereby, p44/p42 MAPK

phosphorylation was time-(Fig.2B) and dose-dependent (Fig.2C) with a significant

increase starting at CEA concentrations of ≥10 ng/ml and reaching its maximum after

15 minutes. Moreover, stimulation of endothelial cells by CEA led to activation of the

phosphoinositide-3-kinase (PI3k)/Akt.pathway (Fig.2D), which represents another

important angiogenic intracellular signaling pathway affecting mainly cell growth,

proliferation, motility and survival (29). Again, Akt activation was time-(Fig.2E) and

dose-dependent (Fig.2F) with maximal Akt phosphorylation occurring at 15 minutes.

In summary, these data suggest that exogenous CEA affects major intracellular

signaling pathways in endothelial cells. 10

When we interfered with these signaling pathways in endothelial cell migration and

proliferation assays, we found that CEA-induced endothelial cell proliferation was

dependent on MEK activity, while Akt-inhibition only partially and not significantly

inhibited CEA-induced endothelial cell proliferation (Fig.2G). CEA-induced endothelial

cell migration, however, was dependent on both pathways, the MEK/ERK- as well as

the PI3k/Akt-pathway, because specific inhibitors these pathways were capable to

block any CEA effect on endothelial cell migration (Fig.2H).

CEA-induced intracellular signal transduction is independent of the VEGF /

VEGFR-1/-2 system

Anti-VEGF/VEGFR system therapies have been introduced for the treatment of

several human cancers. Although this has been proven as successful for patients for

a limited duration of time, a substantial fraction of cancer patients finally becomes

resistant to VEGF-based therapies(30). Therefore, it was important to elucidate

whether VEGF/VEGFR-1/-2 signaling is involved in CEA-induced endothelial cell

activation. CEA-induced p44/p42 MAPK phosphorylation in endothelial cells

remained unaffected by CBO-P11(31), an specific inhibitor of VEGFR-1/-2 (Fig.3A).

Notably, the inhibitory effect of CBO-P11 on VEGF was temporal and partially

reverted after 2 hours (Fig.3B). Consistently, the VEGF-blocking antibody

bevacizumab did not affect CEA-induced p44/p42 MAPK phosphorylation (Fig.3C).

Inhibition of VEGF binding to VEGF receptor-1/2 using CBO-P11 (31) had no effect

on CEA-induced endothelial cell migration, further indicating a VEGF-independent

activity of CEA. Similiar results were obtained by the addition of the functional-

blocking anti-VEGFR-2 antibody (500ng/ml) (Fig.3D). Furthermore, CEA failed to 11

activate VEGFR-2/flk-1 (Fig.3E). These results demonstrate that the VEGF/VEGFR-

1/-2 system seems to be dispensable for CEA-induced endothelial cell activation in

vitro.

Integrin adhesion receptors participate in CEA-induced activation of

endothelial cells

CEA-mediated modulation of integrin-dependent signal transduction has been

reported to occur in tumor cells(14), which appears to be facilitated by co-localization

of CEA and integrins in the same subcellular lipid rafts(15). Therefore, we were

interested to test a potential role of integrins in CEA-induced endothelial cell

activation. Whenever endothelial cells were seeded on integrin-adhesive matrices

such as on the integrin beta-1 or beta-3 ligand fibronectin (Fb), or on the specific

integrin beta-3 ligand fibrinogen (Fg), p44/p42 MAPK became phosphorylated upon

addition of soluble CEA. In contrast, CEA had no impact on p44/p42 MAPK

phosphorylation in endothelial cells seeded on poly-D-lysine (PDL), which does not

interact with integrins, but was phophorylated upon VEGF, indicating endothelial cells

were alive (Fig.4A). The requirement of engaged integrins was further proven by

specific blocking antibodies directed against integrin beta-3 (LM609), which

diminished CEA-induced p44/p42 MAPK phosphorylation (Fig.4B). The direct

involvement of integrin-beta subdomains in CEA-initiated signal transduction was

indicated by the CEA-induced enhancement of activation of critical upstream

components of integrin-signaling, such as FAK, and c-src (Fig.4C and 4D). Since

addition of the MEK-inhibitor U0126 had no effect on CEA-induced FAK-

phosphorylation, FAK appears to signal upstream of p44/42-MAPK (not shown). As a

consequence, integrin-dependent cellular functions such as endothelial cell 12

spreading were enhanced whenever CEA was present (Fig.4E). Importantly, integrin

expression was unaffected by CEA (not shown).

CEA induces integrin activation via inside-out signaling

CEA binding to endothelial cells was shown by immunoprecipitation as well as

immunocytochemistry (Fig.5A and 5B). As a consequence, CEAR was

phosphorylated (Fig.5C), which is consistent with previous observations(32). Integrin

adhesion receptors rapidly increase ligand binding affinity (integrin activation) upon

intracellular stimuli (inside-out signaling). Integrin activation is thereby induced by

talin-binding to cytoplasmic domains of beta-integrins(33). Thus, we have analyzed

talin binding to integrins beta-3 tails using cell lysates from CEA-stimulated or control

endothelial cells using affinity chromatography. We found that talin binding to integrin

beta-3 tails was increased upon CEA-stimulation (Fig.5D), suggesting CEA leads to

integrin activation. Consistently, addition of CEA exerted enhanced binding of both

WOW-1, a Fab antibody fragment specifically recognizing active alphaVbeta3

integrins(20) (Fig.5E) and a FITC-labeled cyclic RGD peptide, a high affinity ligand

for active alphaV-integrins (Supplementary Fig.4A and 4B). The CEA-induced

activated affinity state of integrins was reflected by increased endothelial cell binding

to integrin-adhesive matrix proteins (Fig.5F). Furthermore, we observed that CEA-

induced endothelial cell proliferation was abolished by the blocking anti-beta3 integrin

antibody LM609 when added after cells were adhered (Fig.4F). We conclude that

integrin beta-3 is involved in CEA-induced endothelial cell activation.

CEA overexpression is accompanied by enhanced tumor vascularization

To analyze whether soluble CEA has any biological impact on tumor-angiogenesis,

we next analyzed vascularization in tumor-tissue samples derived from spontaneous

developing gastric tumors in double-transgenic offspring of CEA424-Simian-

Virus40T-antigen (CEA424/Tag) mice mated with human CEA-transgenic

(CEA424/Tag-CEA) mice and compared results with vascular densities of

spontaneous tumors in not CEA-expressing CEA424/Tag-mice. CEA expressing

tumors revealed a markedly higher vascular density than controls (7.04±0.18 blood

vessels/unit ± SEM for CEA424/Tag-CEA vs. 3.43±0.12 in CEA424/Tag; p<0.001)

(Fig.6A). Notably, tumors of same sizes were compared.

To overcome some disadvantages of spontaneous-tumor models such as limitation in

comparing growth dynamic of asynchronous onset of tumor growth, we introduced

tumor transplantation assays. We used a stable CEA-overexpressing sarcoma cell

line (Fig.6B), which were transplanted into BALB/c-mice. When the high CEA-

expressing Meth-A (Meth-A/CEA) sarcomas were compared to the endogenously

non-CEA expressing wild-type sarcomas (Meth-A/wt), a markedly higher tumor

vascularization was observed whenever CEA was present (Fig.6B). Furthermore, the

sizes of explanted Meth-A/CEA tumors were significantly higher than Meth-A/wt

allotransplants. Similar resulst were obtained in a renal cell cancer xenotransplant-

assay using stable-overexpressing CEA (Caki2/CEA) or non-CEA-expressing mock

transfected Caki2 cells (Caki2/mock), which were injected into nude mice

(Supplemental Fig.2B).

To exclude any indirect effects of CEA on angiogenesis such as previous reported

CEA binding to NK-cells via CEACAM1(22), we next performed tumor-transplants in 14

NSG-mice, which are characterized by severe immunodeficiency due to absence of

functional NK-cells, absence of mature T- and B-cells as well as deficiency in

cytokine signaling(34). As expected, growth of tumor-transplants was enhanced in

both groups, but CEA-expressing tumors (MethA/CEA) grew significantly faster than

MethA/wt. Consistently, CEA-expressing tumors explanted after eight days revealed

a higher vascular density (MethA/wt: 63±27.14 blood vessels/unit; MethA/CEA:

139±31.47 blood vessles/unit; n=5 mice per group; p<0.01) and vessel area

(MethA/wt: 1.0±0.23%; MethA/CEA: 3.12±0.54%) compared to MethA/wt. After 21

days MethA/wt-tumors were significant smaller than MethA/CEA-transplants

(MethA/wt: 1.11±0.51g, MethA/CEA 3.24±0.67g; n=5; p<0.01) and tumor growth

curves still diverged. Again, CEA-expressing tumors revealed a higher vessel area

and increased blood vessel density, when compared to controls (Fig.6C).

These data demonstrate that CEA over-expression is accompanied by higher tumor

vascularization.

High CEA expression is a negative prognostic factor for survival of in colorectal

cancer (CRC) patients(27). Thus, we finally analyzed whether CEA-expression

correlates with tumor-angiogenesis in CRC-patients. Immunohistochemical analyzes

of paraffin-embedded-tumor-tissue sections derived from patients with metastasized

CRC, who had either normal (≤5.1 ng/ml) or high (≥34.9 ng/ml) baseline serum CEA-

levels were performed. Thereby, high CEA-expression was accompanied by high

vascularization of the tumor microenvironment represented by an increased vessel

area and an increased number of blood vessels, when compared to tissue samples

derived from patients with low preoperative CEA-levels (Fig.6D, Supporting

Information Table S1). 15

Interference with CEA-Receptor Expression Affects Tumor Vascularization

Finally, we aimed to interfere with CEA-induced tumor-angiogenesis: In a previous

study, CEA was immunological targeted via CEA-mimotope vaccination resulting in a

significantly reduced growth of CEA-expressing tumor transplants (Meth-A/CEA)(16).

Although in CEA-mimotope vaccinated mice tumor-angiogenesis was markedly

reduced (Supplementary Fig.1), we aimed to exclude potential secondary

immunological effects in CEA mimotope-immunized mice including antibody-

dependent cellular cytotoxicity (ADCC) or complement-dependent cytotoxicity (CDC).

Furthermore, targeting integrins is not an adequate model to study CEA-induced

tumor angiogenesis as integrins represent a central player in angiogenesis(35). Thus,

we interfered with CEAR-expression in host endothelial cells in vivo. Notably, MethA-

tumor cells do not express endogenous CEAR, while host endothelial cells express

CEAR, but are deficient in endogenous CEA expression (Supplementary Fig.5H).

In detail, CEAR-expression in mouse endothelial cells was downregulated via a

mouse CEAR-specific shRNA-lentiviral construct. We found that in low CEAR

endothelial cells p44/p42 MAPK-phosphorylation upon CEA stimulation was

decreased, thus, suggesting a critical role for CEAR in CEA-mediated pro-angiogenic

signaling in endothelial cells (Fig.7B).

When we interfered mouse CEAR-expression in a tumor (MethA-) allotransplant

model, we found that the higher tumor vascularization of MethA/CEA tumor

transplants was reverted. Concomitant injection of the mouse CEAR-specific shRNA-

lentiviral construct with Meth-A/CEA tumor cells into nude mice (Meth-A/CEA plus lv)

resulted in a significant reduction of tumor vascularization when compared to Meth-

A/CEA tumors with an empty control virus (Meth-A/CEA) (Fig.7C). The Meth-A/CEA 16

plus lentivirus-transplants had a remarkably reduction of the mean tumor size

(16±24mg) when compared to control virus infected MethA/CEA-transplants

(184±32mg) (Fig.7C). The lentiviral infection rate was equal in both groups (~75+/-

5%) (Supplemental Fig.5D). Similar results were obtained when the shRNA CEAR

lentivirus construct was co-injected to the human Caki2/CEA-tumor transplants (not

shown). Although we cannot exclude any other interaction partners for CEA, targeting

CEAR expression reverted the CEA effect on angiogenesis in these tumor transplant

models. Thus, we conclude that CEA/CEAR-system plays a functional role in tumor

angiogenesis.

Discussion

For the last decade, targeting the VEGF/VEGFR-system was the only anti-angiogenic

approach clinically available to interfere with tumor growth and propagation. While

efficient in certain tumors as such originating from colon, ovary, kidney or brain, not

all cancer patients respond or even become resistant to this anti-angiogenic

approach. In this study, we give first evidence that CEA, which is frequently

overexpressed and released by carcinomas, might also affect pro-angiogenic

endothelial cell behavior.

Initially, we were referred to a potential role of CEA in tumor angiogenesis by a recent

immunological CEA targeting study(16). Thereby, CEA vaccinated mice were found

to have a remarkably lower degree of vascularization in tumors of CEA-immunized

mice, as compared to tumors in control animals (Supplementary Fig.1).

In order to analyze whether CEA directly affected endothelial cells, endothelial cell

behavior was studied in vitro and in vivo. We observed that endothelial cell adhesion,

spreading, migration and proliferation were enhanced in the presence of CEA

(Fig.1A–1D, Fig.2G, Fig.2H, Fig.4E, Fig.4G and Fig.4H). Notably, CEA was active at

concentrations ≥10 ng/ml, which is higher than CEA serum levels found in healthy

individuals (~ ≤5 ng/ml), but is frequently exceeded in patients with certain cancers.

In vitro, CEA-induced endothelial cell activation was independent of the

VEGF/VEGFR-1/-2 system (Fig.3A, Fig.3B and Fig.3D), which might suggest that

CEA bypass VEGF-targeting drugs. However, the clinical relevance remains to be

elucidated and is currently analyzed in a translational approach.

The finding that CEA-induced endothelial cell activation was rather induced via

integrin beta-3 adhesion receptors (Fig.4B and 4F), was consistent with previous

findings that CEA interacts with integrins(14,15). We now observed that CEA led to

an increase in integrin ligand binding activity via enhanced talin/integrin interaction

(Fig.5D). As a consequence, integrin’s most upstream signaling molecules FAK and

c-src became activated (Fig.4A-4D). Consistently, the function-blocking beta-3

integrin antibody LM609 diminished CEA-induced p44/p42 MAPK-phosphorylation

and CEA-induced endothelial cell proliferation. The role of integrins in angiogenesis

is well established and has led to the introduction of interfering agents into pre-clinical

as well as clinical phase I - III trials(36-38).

Notably, endothelial cells do not express CEA (Supplementary Fig.5A-B), but we

demonstrated that soluble CEA interacts with CEAR on endothelial cells (Fig.5A,5B).

Finally, interference with CEAR-expression via siRNA or a shRNA-bearing lentivirus

construct led to a marked reduction of CEA-induced signal transduction in endothelial

cells (Fig.7B) and tumor-angiogenesis (Fig.7C). Further interaction partners,

however, such as the VEGF-interacting proteins neuropilins(39) or other members of

the CEACAM-family(40) can currently not be excluded. Thus, the transmembrane

signaling molecule CEACAM1 has been shown to affect angiogenesis(41,42) and in

NK-cells CEA was described to interact with CEACAM1(22). In experimental settings

we also detected CEACAM1-expression in endothelial cells, but a partial

downregulation of CEACAM1 was dispensable for CEA-induced signaling

(Supplemental Fig.6). This does not exclude a potential prerequisite of low amounts

of CEACAM1 for effective CEA signaling.

To finally analyze the biological relevance of CEA on tumor angiogenesis, we studied

a pre-existing spontaneous stomach tumor model, which developed gastric cancers

with or without transgenic human CEA expression. Furthermore, syngenic and

xenogenic tumor-transplant models using either sarcoma or renal cell cancer cell

lines consistently suggested CEA as an effective stimulus in tumor-angiogenesis,

which was also oberserved in NSG-mice, which lack NK-cell function (Fig.5D). These

data thus suggest that tumor angiogenesis is enhanced by CEA overexpression in

various types of malignancies (Fig.6D). Although preliminary, immunohistochemical

analyzes of human colorectal cancer tissue-sections revealed a correlation between

CEA plasma levels and tumor-angiogenesis, which is currently validated in a larger

prospective translational trial.

In summary, our data derived from in vitro and in vivo experiments demonstrated that

beside its autocrine function to inhibit tumor cell differentiation and apoptosis (10-13,

43,44) soluble CEA expressed and released by adenocarcinomas might bear a

functional role on tumor angiogenesis. This was demonstrated via a transgenic

mouse model of spontaneously developing gastric tumors which expressed human

CEA, two different in vivo tumor transplantation models in mice in three different

genetic backgrounds, three angiogenesis in vivo models and several in vitro models

including cell adhesion, cell spreading, cell migration and cell proliferation and

capillary-like tube formation. Thus, CEA might be considered as a functional active

pro-angiogenic molecule.

Acknowledgments

The work is dedicated to Bernd R. Binder, a main contributor to this work, who

deceased in August 2010.

The project was supported by “Initiative-Krebsforschung”-Foundation (UE71104011)

and FWF-Austrian-Science-Fund (P21301,P23199). Special thanks to Mark

H.Ginsberg (UCSD) for suggesting, supervising and discussing the integrin-

experiments (Fig.4), Matthias Unseld (MUW) for performing the immunoprecipitation

experiments and to Brian Petrich (UCSD) for carefully editing the manuscript.

Author contributions

G.W.P.,C.C.Z.,K.B.: hypotheses generation, manuscript writing. G.W.P.,B.R.B:

designing experiments, data evaluation. M.P.,K.B.,M.U.: performing experiments.

P.U.: in vivo-experiments. W.Z.: transgenic-tumor-tissue-samples, reviewing the

manuscript. F.W.: pathological review of histology, immunohistochemistry.

Conflict of Interest Disclosures The authors declare no competing financial or conflicting interests.

Reference List

(1) Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature

2000;407:249-57.

(2) Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic

switch during tumorigenesis. Cell 1996;86:353-64.

(3) Carmeliet P. Angiogenesis in life, disease and medicine. Nature

2005;438:932-6.

(4) Bais C, Wu X, Yao J, Yang S, Crawford Y, McCutcheon K, et al. PlGF

blockade does not inhibit angiogenesis during primary tumor growth. Cell

2010;141:166-77.

(5) Cassidy J, Clarke S, az-Rubio E, Scheithauer W, Figer A, Wong R, et al.

Randomized phase III study of capecitabine plus oxaliplatin compared with

fluorouracil/folinic acid plus oxaliplatin as first-line therapy for metastatic

colorectal cancer. J Clin Oncol 2008;26:2006-12.

(6) Saltz LB, Clarke S, az-Rubio E, Scheithauer W, Figer A, Wong R, et al.

Bevacizumab in combination with oxaliplatin-based chemotherapy as first-line

therapy in metastatic colorectal cancer: a randomized phase III study. J Clin

Oncol 2008;26:2013-9.

(7) Friedlander M, Brooks PC, Shaffer RW, Kincaid CM, Varner JA, Cheresh DA.

Definition of two angiogenic pathways by distinct alpha v integrins. Science

1995;270:1500-2.

(8) Brooks PC, Montgomery AM, Rosenfeld M, Reisfeld RA, Hu T, Klier G, et al.

Integrin alpha v beta 3 antagonists promote tumor regression by inducing

apoptosis of angiogenic blood vessels. Cell 1994;79:1157-64.

(9) Beauchemin N, Draber P, Dveksler G, Gold P, Gray-Owen S, Grunert F, et al.

Redefined nomenclature for members of the carcinoembryonic antigen family.

Exp Cell Res 1999;252:243-9.

(10) Ordonez C, Screaton RA, Ilantzis C, Stanners CP. Human carcinoembryonic

antigen functions as a general inhibitor of anoikis. Cancer Res 2000;60:3419-

24.

(11) Taheri M, Saragovi HU, Stanners CP. The adhesion and differentiation-

inhibitory activities of the immunoglobulin superfamily member,

carcinoembryonic antigen, can be independently blocked. J Biol Chem

2003;278:14632-9.

(12) Ilantzis C, Demarte L, Screaton RA, Stanners CP. Deregulated expression of

the human tumor marker CEA and CEA family member CEACAM6 disrupts

tissue architecture and blocks colonocyte differentiation. Neoplasia

2002;4:151-63.

(13) Chan CH, Camacho-Leal P, Stanners CP. Colorectal hyperplasia and

dysplasia due to human carcinoembryonic antigen (CEA) family member

expression in transgenic mice. PLoS ONE 2007;2:e1353.

(14) Camacho-Leal P, Zhai AB, Stanners CP. A co-clustering model involving

alpha5beta1 integrin for the biological effects of GPI-anchored human

carcinoembryonic antigen (CEA). J Cell Physiol 2007;211:791-802.

(15) Ordonez C, Zhai AB, Camacho-Leal P, Demarte L, Fan MM, Stanners CP.

GPI-anchored CEA family glycoproteins CEA and CEACAM6 mediate their

biological effects through enhanced integrin alpha5beta1-fibronectin

interaction. J Cell Physiol 2007;210:757-65.

(16) Bramswig KH, Knittelfelder R, Gruber S, Untersmayr E, Riemer AB, Szalai K,

et al. Immunization with mimotopes prevents growth of carcinoembryonic

antigen positive tumors in BALB/c mice. Clin Cancer Res 2007;13:6501-8.

(17) Mechtcheriakova D, Schabbauer G, Lucerna M, Clauss M, De Martin R,

Binder BR, et al. Specificity, diversity, and convergence in VEGF and TNF-

alpha signaling events leading to tissue factor up-regulation via EGR-1 in

endothelial cells. FASEB J 2001;15:230-42.

(18) Koshelnick Y, Ehart M, Hufnagl P, Heinrich PC, Binder BR. Urokinase receptor

is associated with the components of the JAK1/STAT1 signaling pathway and

leads to activation of this pathway upon receptor clustering in the human

kidney epithelial tumor cell line TCL-598. J Biol Chem 1997;272:28563-7.

(19) Kiosses WB, Shattil SJ, Pampori N, Schwartz MA. Rac recruits high-affinity

integrin alphavbeta3 to lamellipodia in endothelial cell migration. Nat Cell Biol

2001;3:316-20.

(20) Pampori N, Hato T, Stupack DG, Aidoudi S, Cheresh DA, Nemerow GR, et al.

Mechanisms and consequences of affinity modulation of integrin

alpha(V)beta(3) detected with a novel patch-engineered monovalent ligand. J

Biol Chem 1999;274:21609-16.

(21) Prager GW, Poettler M, Schmidinger M, Mazal PR, Susani M, Zielinski CC, et

al. CD98hc (SLC3A2), a novel marker in renal cell cancer. Eur J Clin Invest

2009;39:304-10.

(22) Stern N, Markel G, Arnon TI, Gruda R, Wong H, Gray-Owen SD, et al.

Carcinoembryonic antigen (CEA) inhibits NK killing via interaction with CEA-

related cell adhesion molecule 1. J Immunol 2005;174:6692-701.

(23) Nockel J, van den Engel NK, Winter H, Hatz RA, Zimmermann W, Kammerer

R. Characterization of gastric adenocarcinoma cell lines established from

CEA424/SV40 T antigen-transgenic mice with or without a human CEA

transgene. BMC Cancer 2006;6:57.

(24) Eades-Perner AM, van der Putten H, Hirth A, Thompson J, Neumaier M, von

KS, et al. Mice transgenic for the human carcinoembryonic antigen gene

maintain its spatiotemporal expression pattern. Cancer Res 1994;54:4169-76.

(25) Thompson J, Epting T, Schwarzkopf G, Singhofen A, Eades-Perner AM, van

Der PH, et al. A transgenic mouse line that develops early-onset invasive

gastric carcinoma provides a model for carcinoembryonic antigen-targeted

tumor therapy. Int J Cancer 2000;86:863-9.

(26) Liu S, Thomas SM, Woodside DG, Rose DM, Kiosses WB, Pfaff M, et al.

Binding of paxillin to alpha4 integrins modifies integrin-dependent biological

responses. Nature 1999;402:676-81.

(27) Pedrazzani C, Cerullo G, Marrelli D, Fernandes E, Carlucci F, Corso G, et al.

Is circulating D-dimer level a better prognostic indicator than CEA in

resectable colorectal cancer? Our experience on 199 cases. Int J Biol Markers

2010;25:171-6.

(28) Johnson GL, Lapadat R. Mitogen-activated protein kinase pathways mediated

by ERK, JNK, and p38 protein kinases. Science 2002;298:1911-2.

(29) Fine B, Hodakoski C, Koujak S, Su T, Saal LH, Maurer M, et al. Activation of

the PI3K pathway in cancer through inhibition of PTEN by exchange factor P-

REX2a. Science 2009;325:1261-5.

(30) Fischer C, Mazzone M, Jonckx B, Carmeliet P. FLT1 and its ligands VEGFB

and PlGF: drug targets for anti-angiogenic therapy? Nat Rev Cancer

2008;8:942-56.

(31) Zilberberg L, Shinkaruk S, Lequin O, Rousseau B, Hagedorn M, Costa F, et al.

Structure and inhibitory effects on angiogenesis and tumor development of a

new vascular endothelial growth inhibitor. J Biol Chem 2003;278:35564-73.

(32) Laguinge L, Bajenova O, Bowden E, Sayyah J, Thomas P, Juhl H. Surface

expression and CEA binding of hnRNP M4 protein in HT29 colon cancer cells.

Anticancer Res 2005;25:23-31.

(33) Shattil SJ, Kim C, Ginsberg MH. The final steps of integrin activation: the end

game. Nat Rev Mol Cell Biol 2010;11:288-300.

(34) Liu X, Lai L, Wang X, Xue L, Leora S, Wu J, et al. Ribonucleotide reductase

small subunit M2B prognoses better survival in colorectal cancer. Cancer Res

2011;71:3202-13.

(35) Scaringi C, Minniti G, Caporello P, Enrici RM. Integrin inhibitor cilengitide for

the treatment of glioblastoma: a brief overview of current clinical results.

Anticancer Res 2012;32:4213-23.

(36) Hynes RO. A reevaluation of integrins as regulators of angiogenesis. Nat Med

2002;8:918-21.

(37) Hersey P, Sosman J, O'Day S, Richards J, Bedikian A, Gonzalez R, et al. A

randomized phase 2 study of etaracizumab, a monoclonal antibody against

integrin alpha(v)beta(3), + or - dacarbazine in patients with stage IV metastatic

melanoma. Cancer 2010;116:1526-34.

(38) Desgrosellier JS, Cheresh DA. Integrins in cancer: biological implications and

therapeutic opportunities. Nat Rev Cancer 2010;10:9-22.

(39) Geretti E, Shimizu A, Klagsbrun M. Neuropilin structure governs VEGF and

semaphorin binding and regulates angiogenesis. Angiogenesis 2008;11:31-9.

(40) Oliveira-Ferrer L, Tilki D, Ziegeler G, Hauschild J, Loges S, Irmak S, et al.

Dual role of carcinoembryonic antigen-related cell adhesion molecule 1 in

angiogenesis and invasion of human urinary bladder cancer. Cancer Res

2004;64:8932-8.

(41) Gerstel D, Wegwitz F, Jannasch K, Ludewig P, Scheike K, Alves F, et al.

CEACAM1 creates a pro-angiogenic tumor microenvironment that supports

tumor vessel maturation. Oncogene 2011;30:4275-88.

(42) Kilic N, Oliveira-Ferrer L, Wurmbach JH, Loges S, Chalajour F, Neshat-Vahid

S, et al. Pro-angiogenic signaling by the endothelial presence of CEACAM1. J

Biol Chem 2005;280:2361-9.

(43) Samara RN, Laguinge LM, Jessup JM. Carcinoembryonic antigen inhibits

anoikis in colorectal carcinoma cells by interfering with TRAIL-R2 (DR5)

signaling. Cancer Res 2007;67:4774-82.

(44) Soeth E, Wirth T, List HJ, Kumbhani S, Petersen A, Neumaier M, et al.

Controlled ribozyme targeting demonstrates an antiapoptotic effect of

carcinoembryonic antigen in HT29 colon cancer cells. Clin Cancer Res

2001;7:2022-30.

Figure Legend

Figure 1 CEA directly affects endothelial cell behavior in vitro and

angiogenesis in vivo.

(A) Endothelial cell proliferation assay: CEA and VEGF at concentrations indicated;

24 hours. n=3 (B) Endothelial cell proliferation-assay: exogenous CEA: 100 ng/ml

VEGF: 10 ng/ml, n=3). CEA plus VEGF vs. CEA alone: p=0.04 (T-test); 24 hours; (C)

Transwell endothelial cells migration-assay: control (1% BSA): 269 ± 16 cells / unit,

exogenous CEA (concentrations as indicated); 4 hours led to an enhanced trans-

migratory response compared to control. Absolute numbers transmigrated endothelial

cells per unit are indicated (n=3). (D) Aortic ring assay: aortic rings cultured in growth

factor-reduced matrigel in the presence of exogenous CEA (100 ng/ml) or VEGF (30

ng/ml) after six days; MethA/CEA supernatant (adjusted to 100 ng/ml) was incubated

with anti-CEA beads of IgG-beads; MethA/wt: CEA: 0 ng/ml or control: 3% BSA.

Quantification is given in bars; error bars: SD, n=3. (E) Directed in vivo angiogenesis

assays (DIVAA™): CEA (1 µg/ml) or BSA 1% was embedded into growth-factor-

reduced basement membrane-matrix of tubes, implanted into Bl6 mice for 11 days.

Endothelial cells: anti-CD31 stain (green); perfused vessels: Hoechst 33342 (blue).

Quantification: vessel area (left bars), perfusion area (right bars). n=8 tubes/group.

error bars: SEM;

*p<0.05 (A-E)

Figure 2. CEA induces intracellular signal transduction in endothelial cells.

(A-F): Western blots from HUMECs whole-cell lysates stimulated with CEA (100

ng/ml or as indicated) or VEGF(10 ng/ml) for 15 min (unless otherwise indicated).

Mean ± SEM phosphorylation : pan-protein ratio is given by the bars; n≥3

experiments. (A-C) phosphorylation of p44/42 MAPK by CEA was (A) blocked by

MEK-inhibitor UO126 (1 µg/ml), (B) was time- and (C) was dose-dependent. Akt

Ser473 phosphorylation by CEA was (D) blocked by the PI3k-inhibitor LY294002

(10µM), (E) was time and (F) dose dependent. (G) HMVEC proliferation assay:

VEGF: 10 ng/ml; CEA: 100 ng/ml; MEK-inhibitor PD98059 (PD): 10µM; PI3K-

inhibitor: Ly294002 (Ly): 10µM. n=3. (H) HMVEC migration assay: VEGF: 10 ng/ml;

CEA: 100 ng/ml; PD98059 (PD): 10µM; Ly294002 (Ly): 10µM. n=3.

* p<0.05 (Dunnett’s test).

Figure 3 CEA acts independently of the VEGF / VEGFR-1/-2 system

(A-C): Western blots for p44/p42 MAPK from HUMECs whole cell lysates; bars

represent densitometry quantification of phospho-p44/p42 versus pan-p44/p42 ratio:

mean ± SEM of n≥3 experiments; *p<0.05: (A) CEA: 100 ng/ml, VEGF: 10 ng/ml,

CBO-P11: 40 µg/ml (CBO); 15 min (B) the CBO-P11 (40 µg/ml; CBO) effect was

time-dependent with a partial reduction of the inhibitory effect at 120 and 240 min.

Loading control: anti-glyceraldehyde-3-phosphate-dehydrogenase (C)

Immunoprecipitation of VEGF before incubation of HUMECs with CEA (100 ng/ml) or

VEGF (10 ng/ml) prevented p44/p42 MAPK phosphorylation upon VEGF, but had no

effect on CEA. (D) CEA-(100 ng/ml) induced HUMECs transmigration was neither

affected by CBO-P11 (CBO, 40 µg/ml) (left panel), nor by anti-VEGFR-2 antibody

(500 ng/ml) (right panel); (n=4, *p<0.05, #p>0.05) (E) Western blot: VEGF-R Tyr1175

phosphorylation upon VEGF – but not CEA (100ng/ml) was blocked by the addition

of CBO-P11 (40 µg/ml; CBO).

Figure 4 Endothelial cell activation by CEA depends on functional active beta-3

integrins. (A-D) Western blots: (A) p44/p42 MAPK phosphorylation upon CEA (100

ng/ml) or VEGF (10 ng/ml): Endothelial cells were seeded on the integrin ß1- or ß3-

ligand fibronectin (Fb) or the specific ß3-ligand fibrinogen (Fg), or the non-integrin-

ligand poly–D–lysine (PDL); 10min (B) Endothelial cells were seeded on Fg before

they were incubated either with VEGF (10 ng/ml) or CEA (100 ng/ml) in the presence

of the beta3 integrin-blocking antibody LM609 or mouse-IgG (control). (C) Y576 FAK

phosphorylation in HUMECs seeded on fibronectin (Fb), fibrinogen (Fg) or Poly-D-

lysine (PDL) in the presence or absence of CEA (100 ng/ml) or VEGF (10 ng/ml), 10

min. (D) Src Tyr416-phosphorylation (represents active c-src) in HUMECs upon CEA

involves integrin engagement. (E) HUMEC spreading assays: cell surface areas were

measured after 10min adhesion to Fb or after 30min adhesion to Fg. Control: 1%

BSA; VEGF: 10ng/ml; CEA: 100 ng/ml. (F) HUMEC-proliferation assay: integrin beta-

3 blocking antibody LM609 (1µg/ml) abolished the proliferative response towards

CEA (100 ng/ml). (A – F) n≥3 independent experiments; HUMECs were used.

*p<0.05, #p>0.05.

Figure 5 CEA induces integrin beta-3 activation

(A) Co – immunoprecipitation: lysates of mouse endothelial cells, which were

stimulated either with 500ng/ml human CEA (lane 2 and 3) or with BSA 1% (control,

lane 1). Immunoprecipitation with mAB anti-CEAR beads (lane 1 and 2) or IgG-beads

(lane 3); Western blots: anti-CEA. n=4 (B) Confocal laser-scanning-microscopy of

MUMECs stimulated with CEA (100ng/ml) revealed co-localization of CEA (green)

and CEAR (red); DAPI (blue); n=4 (C) Immunoprecipitation of CEAR: CEA

(100ng/ml, 10min) led to an increase in CEAR-tyrosine phosphorylation. (D) Affinity

chromatography: immobilized Neutravidin beads were loaded with β3(HisAvi-B3) or

randomized-sequence-β3 (HisAvi-B3rand) tail proteins. Bound proteins from CEA-

activated HUMEC- or control HUMEC-lysates were separated on 4-20% SDS-PAGE

under reducing conditions; Western blots: anti-talin. Equal loading (Loading) was

assessed by Coomassie Blue-staining of protein bound beads. n=5. (E) FACS

analysis: WOW-1 binding to active integrin-β3: CEA (100ng/ml, 30min) increased the

activate state of αVβ3 integrins; Mn++-stimulation: positive control. Blank: mouse-IgG

(F) Adhesion assay: Endothelial cell stimulated by CEA induced enhanced adhesion

to Fg (black bars) or FN (white bars) when compared to control (BSA 1%), 30min

(n=3).

Figure 6 CEA overexpression is accompanied by increased tumor

angiogenesis.

(A) Spontaneous transgenic tumor model: Immunohistochemical staining for

endomucin (upper panels), human CEA (middle panels) or haematoxylin/eosin-(HE)

(lower panels) in tumor tissue samples derived from spontaneous gastric tumors of

transgenic mice expressing human CEA (CEA424/Tag-CEA) or not (CEA424/Tag):

Tumors of same size were analyzed for angiogenesis: CEA-expressing tumors

revealed a markedly higher vascular density than controls. Size bars: 50µm; n=8;

*p<0.05 (B) Allotransplant model: Immunohistochemical staining for endomucin in

MethA/wt and MethA/CEA tumor-tissue sections revealed an increase in blood

vessels per unit, tumor vessel area and tumor weight (day 8) whenever CEA was

present. Tumor growth over time (d = days) was calculated in grams (g). (*p<0.05;

error bars= SEM; number of microscopic fields analyzed= 8, 4 mice/ group, 1 unit=

1mm2, size bars: 100 µm). (C) Allotransplants of MethA/wt or MethA/CEA cells into

NSG-mice: CEA-expression is accompanied by enhanced tumor growth independent

of NK-cell activity. After long term (21d) incubation tumor vascularization per field

(1mm²) was significantly higher in CEA-expressing tumors than in controls. *p<0.05;

error bars= SEM; 5 mice/ group. (D) Immunohistochemical analysis of tumor-tissue

sections derived from mCRC-patients revealed a direct correlation between serum

CEA levels and tumor vascularization: In patients with normal serum CEA levels (low

sCEA: ≤5.1 ng/ml) tumor vascular densities and vessel areas were lower than in

patients with high serum CEA levels (high sCEA: ≥34.9 ng/ml). Representative

pictures are displayed. Panels illustrate quantification of tumor vascularization; mean

+/- SEM. field= 0.5mm2, size bar: 100µm, n=24. *p<0.01

Figure 7 CEAR is involved in CEA-induced angiogenesis

(A) Western blots: downregulation of CEAR in MUMECs by shRNA-lentiviral

constructs: ≥ 90% downregulation after 48 hours. (B) Western blot for p44/p42

MAPK-phosphorylation in MUMEC-lysates: Transient CEAR downregulation (~10%

of control) by siRNA led to diminished phosphorylation of upon CEA (100 ng/ml), but

did not VEGF. n=3 (C) In vivo tumor allotransplant assay in nude mice:

immunohistochemistry for endomucin (brown) in Meth-A/wt (control), Meth-A/CEA or

Meth-A/CEA plus shRNA lentiviral (+lv) on day 8; Calculated tumor growth curve (d=

days) (p<0.05; error bars= SEM; n=18, 6 mice/group). MethA/CEA transplants

showed increased blood vessels/unit as well as vessel area. Size bars: 100µm;

*p<0.05

Soluble Carcinoembryonic Antigen (CEA) Induces Endothelial Cell Activation and (Tumor-)Angiogenesis

Kira H. Braemswig, Marina Poettler, Matthias Unseld, et al.

Cancer Res Published OnlineFirst October 11, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-13-0123

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2013/10/11/0008-5472.CAN-13-0123.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet been Manuscript edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2013/10/10/0008-5472.CAN-13-0123. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.