[CANCER RESEARCH57, 4432-4436, October I, 1997]

cally to colonize the liver (13, 14). This effect occurs with CRC cellsthat neither produce CEA nor bind to CEA on a solid phase in vitro(13, 14). Therefore, CEA may facilitate metastasis indirectly througha mechanism that is independent of its homotypic adhesive properties.

TNF-a, IL-la, and IL-6 are proinflammatory cytokines involved intumorigenesis and metastasis as well as in the inflammatory response(15—17).Patients with advanced human CRCs have elevated systemicserum IL-6 but not IL-la or TNF-a levels (18). IL-6 also acts as ahepatocyte-stimulating factor to induce acute-phase proteins in hepatocytes (19). We have reported that in vitro treatment of Kupifer cellswith CEA results in the production of IL-6, TNF-a, IL-la, and IL-1(3in rats and humans (20). LL-6 along with TNF-a and IL-la increasesthe expression of intercellular adhesion molecule 1 and endothelialleukocyte adhesion molecule 1 on human endothelial cells (18). Wehypothesized that CEA in the circulation stimulates Kupffer cells toinduce a systemic cytokine response and ultimately enhance experimental metastasis.

The present study compared the ability of systemic CEA and LPSto stimulate serum levels of IL-6, IL-la, and TNF-a in mice andexamined the role of tyrosine phosphorylation in this response.Groupsof mice wereinjectedi.v. with CEA with or without pretreatment with genistein, a tyrosine kinase inhibitor. In a similar manner,groups of mice were also injected i.v. with asialoCEA or PELPKconjugated albumin (the pentapeptide sequence of CEA that is aligand for the CEA receptor on Kupffer cells; Refs. 21 and 22) todetermine whether the induction of the cytokines was a specificresponse to CEA. Levels of cytokines in the circulation were determined by ELISA. Tyrosine-phosphorylated proteins were identified inmouse Kupffer cells that had been isolated and exposed to CEA andLPS with or without genistein. The results suggest that i.v. CEAstimulates murine Kupffer cells to produce more IL-6 than TNF-a orIL-la through transduction pathways that may be different from thoseinduced by LPS.

MATERIALS AND METHODS

Animals. Four- to 5-week-old athymic nude and BALB/c mice weighing30—40g each were obtained from Harlan Sprague Dawley Inc. (Indianapolis,IN) and housed five mice per cage under specific pathogen-free conditions.The standard diet and water were given ad libitum. All operative procedureswere performed under general anesthesia using i.p. administration of 50 mg/kgsodium pentobarbital from Abbott Laboratories (North Chicago, IL). Animal

experiments were approved by the Beth Israel Deaconess Medical Center,West Campus Animal Care and Use Committee and conducted in accordancewith the guidelines issued by the NIH for the care of laboratory animals.

Reagents. CEA was prepared from hepatic metastases of CRC patients byperchloric acid extraction followed by chromatography. The purity of thepreparations was determined by SDS-PAGE, high-performance liquid chromatography analysis and by immunoreactivity as described previously (23).Endotoxin levels were measured to be <0.1 ng per 1 .sgof CEA. AsialoCEAwas prepared by treatment with neuraminidase (Vibrio cholerae; Ref. 24).AAG was isolated from pooled human serum (25). LPS from Salmonellaminnesota was obtained from Sigma Chemical Co. (St. Louis, MO), and

endotoxin-free PBS was obtained from Life Technologies, Inc. (Grand Island,

NY). GBSS was supplied by Life Technologies, Inc. Genistein (4',5,7-trihy

droxyisoflavone) was obtained from Dr. V. Steele of the Chemoprevention

Program, National Cancer Institute (Bethesda, MD).

ABSTRACT

Carcinoembryonic antigen (CEA) may promote experimental metasta

sis through production of cytokines. The effect of systemic CEA on theproduction of proinflammatory cytokines was investigated in mice and

compared to levels induced by lipopolysaccharide (LPS). Serum concentrations of interleukin (IL)-6 peaked 1 h after an i.v. CEA injection of 40gLglmouse to 37—54%of the maximal level induced by a 1 @.sg/mouseinjection of LPS in both normal and immunoincompetent mice. The CEA

inductionof IL-6 wasa specificresponse,becausethe peptidePELPK (thepentapeptide on CEA that is the ligand for the CEA receptor on Kupifercells) conjugated to albumin induced 30% of the maximal CEA responsefor IL-6, whereas the specificity control PELGK-conjugated albumin didnot. IL-la and tumor necrosis factor (TNF)-a levels after i.v. injection ofCEA were only 3-5% ofthose induced by LPS. The IL-6 responses of micepretreated with 100 @&g/kggenistein were decreased by more than 40%.However, genistein inhibited the TNF-a response to LPS by 46% butincreased the CEA-induced response by 300%. When murine Kupffercells were stimulated with LPS or CEA in vitro, LPS increased tyrosinephosphorylation of a M@30,000 protein, whereas CEA decreased phosphorylation of a M@ 60,000 protein and did not increase phosphorylation

ofthe Mr @°@°°°protein. Thus, i.v. CEA stimulates production of IL-6 andTNF-a after binding to Kupffer cells through signal transduction pathways that appear to be different from those stimulated by LPS.

INTRODUCTION

CEA4 is a Mr 180,000200,000 glycoprotein tumor marker identifled originally by Gold and Freeman (1). Preoperative serum levels ofCEA in the circulation of greater than S ng/ml are associated withsignificantly shorter disease-free survival in Dukes' stage H—Illpatients (2, 3), and elevations in postoperative serum CEA concentrations are also associated with disease progression (4). Expression ofCEA in weakly metastatic human CRC cells that lack CEA increasesthe number of liver colonies after intrasplenic injection into nude mice(5, 6). Nevertheless, the exact role of CEA in metastasis is not yetcompletely understood. CEA, nonspecific cross-reacting antigen, thebiliary glycoproteins, and the pregnancy-specific glycoproteins form asingle gene family located on chromosome 19 (7) that are part of theimmunoglobulin supergene family that includes neural cell adhesionmolecule (8) and other adhesion molecules. CEA may act as anintercellular adhesion molecule to promote CRC cell attachment atmetastatic sites (9, 10) or as a ligand for selectins (11) and galectin-3(12). Systemic pretreatment of athymic nude mice with CEA increasesthe ability of weakly metastatic human CRC cells injected intraspleni

Received 4/22/97; accepted 8/2/97.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance withI8 U.S.C. Section 1734 solely to indicate this fact.

@ This work was supported by National Cancer Institute Grants CA42857 andCA44704 (to J. M. J.) and by National Aeronautics and Space Administration GrantsNAG 9-650 (to J. M. J.) and T32-CA 9850 (to K. H. E.).

2 Present address: Virginia Surgical Associates. Fairfax. VA 22033.

3 To whom requests for reprints should be addressed, at Department of Surgery,

University of Pittsburgh Medical Center, 300 Liliane Kaufmann, 200 Lothrop St.,Pittsburgh, PA 15213. Phone: (412) 692-2957; Fax: (412) 692-2520; E-mail:jessupj@delphi.com.

4 The abbreviations used are: CEA. carcinoembryonic antigen; AAG, al acid glyco

protein; CRC, colorectal carcinoma; IL, interleukin; GBSS, Gey's balanced salt solution;LPS, lipopolysaccharide; TBS, Tris-buffered saline; TBS-T, TBS-0.l% Tween; TNF-a,tumor necrosis factor-a.

4432

In Vivo Induction of Murine Cytokine Production by Carcinoembryomc Antigen1

Kirsten H. Edmiston,2 Aniruddha Gangopadhyay, Yutaka Shoji, Alex P. Nachman, Peter Thomas, andJ. Milburn Jessup3

Laboratory of Cancer Biology. Department of Surgery, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts 02215

on April 17, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CARCINOEMBRYONICANTIGEN INDUCTION OF CYTOKINES

Two pentapeptides (PELPK and PELGK) were synthesized on the basis ofthe CEA-binding sequence PELPK (Tyr-Pro-Glu-Ile-Pro-Lys; amino acids

108—112of CEA) as described previously (22). These peptides were thenconjugated with hydroxysuccinimidyl 4-azidobenzoate (Sigma) and coupled to

BSA by exposure to short-wave UV radiation for 30 mm at room temperature.The reaction gives a maximum incorporation of 10 mol of peptide/mol ofalbumin (22).

Serum Cytokine Determination. Groupsof two to three mice were injected i.v. with 10—40pg/mouse of CEA, 40 gtg/mouse asialoCEA, 40

@tg/mouseAAG, 1 @tg/mouseLPS, 40 @tg/mousePELPK or PELGK-conjugated albumin, or 0.1 ml of PBS alone. Other mice were pretreated with 100mg/kg genistein dissolved in 0.1% ethanol-PBS injected i.p. 2 h prior to CEA

or LPS injection. Control experiments were performed using 0.1% ethanolPBS alone. At 10 mm to 24 h after injection, approximately 1 ml of blood wasobtained by cardiac puncture. Blood samples were centrifuged at 14,000 rpmfor 10 mm at 4°Cto separate the serum. Concentrations of IL-a, IL-6, andTNF-a in the serum samples were determined by ELISA using commerciallyavailable kits from Endogen Corp. (Cambridge, MA) and following the recommended protocol of the manufacturer. Quantification of secreted cytokineswas accomplished by normalization of the ELISA data with a standard cyto

kine dose curve. Cytokine levels for each time point were measured in two to

four independent experiments. There was no cross-reactivity between CEA andany of the cytokine ELISA assay kits.

KupfTer Cell Isolation. Mouse Kupifer cells were isolated using a modifled protocol of collagenase infusion followed by metrizamide density gradient

separation with selective adherence to plastic (24). Briefly, the mouse liver wasperfused through the portal vein with Ca2@-and Mg2@-freebuffer containing0.83% sodium chloride, 0.05% potassium chloride, and 0.005 @M(2-hydroxymethyl)piperazine-N-1-(2-ethanesulfonic acid), followed by perfusionwith 0.05% collagenase (0.025% type II and 0.025% type IV) in GBSS at a

flow rate of 5 mI/mm for 2 mm. The liver was then excised, incubated for 30mm at 37°C in 5% CO2. minced, and passed through a 150-gm steel grid. The

cells were resuspended in GBSS and allowed to settle for 1 mm. The pellet

containing hepatocytes was discarded, and the supernatant was centrifuged at300 X g for 10 mm at 25°C.Kupifer cells were separated from this cell pelletby isopyknic sedimentation using an 18% metrizamide gradient in GBSS

without NaC1, overlaid with 1.0 ml GBSS with NaCI and centrifuged at1264 X g at the interface for 20 mm at 20°C using an SW-4lTi swing

rotor/L8-M ultracentrifuge from Beckman Instruments (Palo Alto, CA). Cellswere aspirated from the interface and incubated for 20 mm at 37°C in T25

culture flasks. Nonadherent cells were discarded by washing, and the attachedwere used for the experiments. These cells were cultured in RPM! 1640 with

100 units/mi penicillin G and 100 pg/mi streptomycin, 1% L-glutamlne, and10% fetal bovine serum in a humidified 5% CO2 environment at 37°C.

Identification of Tyrosine Phosphoproteins. Cultures of Kupffer cellswere incubated with 40 @g/mlCEA or AAG, 1 @sg/mlLPS in RPM! 1640, or0.1 ml of medium alone. Another set of Kupffer cells was pretreated with 74jLM of genistein (26) dissolved in 0.1% ethanol in complete RPM! for 2 h prior

to treatment with CEA, AAG, or LPS. Control experiments were performedusing 0.1% ethanol alone in complete RPMI. After a 1-h incubation at 37°C,2 X iO@Kupifer cells were lysed with a lysis buffer containing 0.5 mM NaVO4,50 @tg/mlL-1-p-tosylamlnO-2-phenylethylchloromethyl ketone, 50 @.sg/mltosyl lysine chloromethyl ketone, 10 @tg/miaprotinin, 0.2 m@iphenylmethylsulfonyl fluoride, 10 @.sg/mlleupeptin, 10% n-octyl glucoside, and 50 @tg/m1DNase in 0.1 MTBS. Precipitate was removed by centrifugation at 3000 rpmfor 10 mm at 4°C.Protein concentrations in the lysates were determined, and50-,.Lg samples were resolved on a 10% SDS-PAGE under reducing conditionsand transferred to a nitrocellulose membrane. After blocking for 16 h at 23°Cin TBS with 10% skim milk, the blot was incubated for 1 h at room temper

ature with antiphosphotyrosine antibody (PY-20; Transduction Laboratory,Lexington, KY) diluted 1:1000 in TBS-T with 10% skim milk. Blots werewashed with TBS-T and incubated for 1 h at room temperature with horse

radish peroxidase-conjugated sheep antimouse IgGs from Amersham LifeScience (Buckinghamshire, United Kingdom), diluted 1:1000 in TBS-T in10% skim milk, and developed on KOdakX-Omax film (Kodak, Rochester,NY) using an enhancedchemiluminescencesystem (Amersham).

Statistical Analysis. The results of the ELISA assays are expressed asmean ± SE, with differences among groups of means tested by one-way

ANOVA. All tests were performed in at least duplicates. When ANOVA

demonstrated that means within an experiment were significantly different

from one another, significance between individual group means was tested

with the Fisher PSLD with a significance level of 5%. All calculations were

performed on a Macintosh microcomputer using Stat View SE + Graphics

(Abacus Concepts, Berkeley, CA).

RESULTS

Systemic Cytokine Production. The ability of systemic CEA toinduce cytokine production was first analyzed in athymic nude mice.IL-6 levels began to increase 10 mm after the injection of 40 @tgofCEA and peaked after 1 h at 1944 ±55 pg/ml. By comparison, 1 @sgof LPS stimulated a 2-fold greater peak of IL-6 1.5—2h after injection,whereas 40 i.@gof AAG stimulated a peak of IL-6 that was one-sixththat of CEA 4 h after injection (Fig. 1; Table 1). CEA induced IL-6production in a dose-dependent manner with doses greater than 10 @tg

needed to induce measurable levels of IL-6 in the serum (Fig. 2).Interestingly, the rise in serum IL-6 levels parallels the enhancementof liver colony formation observed with the weakly metastatic KM12c human CRC cell line after CEA injection (Fig. 2). The levels of

IL-6 in the peripheral blood returned to baseline by 8 h (Fig. 1).CEA binds to a Mr 80,000 binding protein expressed on Kupffer

cells and alveolar macrophages through the pentapeptide PELPK (23).

AsialoCEA, which is identical to CEA but without sialic acid residues, is taken up by hepatocytes rather than Kupffer cells because ofthe greater affinity of the hepatocyte asialoglycoprotein receptor forasialoCEA (24). Serum IL-6 concentrations did not increase 1 h afterinjection of 40 @gof asialoCEA (Fig. 3). However, the injection of 40;.tg of PELPK-conjugated albumin induced serum IL-6 responses thatwere 31% of the maximal response to 40 .tg of native CEA (Fig. 3).In contrast, 40 i@gof PELGK-conjugated albumin, which does notbind to the Mr 80,000 CEA binding receptor, induced a minor IL-6response (Fig. 3). Thus, both CEA and the pentapeptide sequence(PELPK), which bind to the CEA receptor on Kupffer cells, induce anIL-6 response in athymic nude mice.

CEA did not induce a strong systemic IL-la response in athymicnude mice. Peripheral blood serum levels of IL-la reached 19.6 ±1.2pg/mi after CEA injection, which was not significantly different from

that seen with PBS alone (1 1.1 ±3.7 pg/mI at 2—4h). AAG also didnot significantly stimulate IL- 1a production over baseline. In contrast,

4000

@@1

EDl0@

Co-I

0

3000

2000

1000

0 6 12 18 24

Time (hr)Fig. 1. IL-6 expression in athymic nude mice treated with 40 @sgof AAG (•),40 @.sg

of CEA (R), or 1 @gof LPS ( •) in endotoxin-free PBS. Four- to 5-week-old athymicnude mice were injected iv. through the dorsal tail vein. Blood was collected at the timesindicated, and serum was assayed for IL-6 concentrations by ELISA. Values were plottedafter deducting the untreated control values. Data are means; bars, SE.

4433

on April 17, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Table 1 Summaryof cytokine responsesto CEA aad relatedsubstances―Nude

miceBALB/c mice

Peak TimetoPeakTimetoCytokineAgentresponse(%)peak (h)response (%) peak(h)IL-6PBS

LPSAsialoCEACEAAAG39

13610 —

10 0.31944 54602 171

21180

—398 — 2ND1329 37 1319 94IL-laPBS

LPSAsialoCEA

CEA4

1404 —

13 320 51

I113

2 1131 — 1

ND5 41TNF-aPBS

LPSAsialoCEA

CEA9

0.24988 —

3 0.1156 31

2110

4882 — 1ND111 2 1

10

% of Mice‘60 WIth

Metastases.4@ (I)

CARCINOEMBRYONICANTIGEN INDUCTION OF CYTOKINES

the IL-6 responses to CEA (40 @.tg/mouse)and LPS (1 @tg/mouse)by46% and 47%, respectively (Fig. 4A). In contrast, pretreatment withgenistein induced a 300% increase in the TNF-a response to CEA but an

81% decrease in TNF-cr response to LPS (Fig. 4B). The IL-la responseto CEA was not affected significantly by pretreatment with genistein(data not shown). These results suggest that CEA and LPS affect cytokineproduction through different signal transduction pathways.

Effect of CEA on Tyrosine Phosphorylation of Murine KupiferCell Proteins. Because genistein pretreatment had different effectson the productionof IL-6 and TNF-cs, the patternof tyrosine phosphorylation was next assessed in isolated Kupffer cells incubated withCEA and LPS in vitro. Two proteins (Mr 60'000 and 80,000) werephosphorylated in control Kupifer cells. After 1 h, LPS stimulatedtyrosine phosphorylation of a Mr 30,00035,000 protein. After a 1-hexposure to CEA, the tyrosine phosphorylation of a Mr 80,000 proteinwas increased, whereas the phosphoiylation of a Mr 60'000 @)I@Oteifl5wasdecreased (Fig. 5). Thus, the signal transduction pathways by which CEAinduces IL-6 and TNF-a appear to be different from those that LPS uses.

The present study seeks to explain observations made by us andothers regarding the role of CEA in metastasis. Previously, we haveshown that 40 @tgof CEA injected i.v. into athymic nude miceincreases the number of hepatic metastases produced by three weaklymetastatic human CRC cell lines (KM-12c, MIP-lOl, and Clone A)but not by either highly metastatic cell lines or a tumorigenic nonmetastatic cell line (10, 13, 14). In vitro, KM-12c cells bind to CEAattached to a solid phase by homotypic mechanisms, whereas MIP

101 and Clone A cells do not (29). These data are consistent with CEAfacilitating increased retention and survival of weakly metastatictumor cells in liver by a mechanism that does not involve directhomotypic adhesion between CEA molecules on the tumor cells andCEA-related epitopes within the hepatic microcirculation. We hypothesized, therefore, that systemic pretreatment with CEA could induce ahost cytokine response that modulates the tumor cell-environmentinteraction to favor survival of metastatic cells.

Mouse Kupifer cells produce various cytokines, including IL-i a,IL-6, TNF-a, and WN-a/f3 in response to LPS (30). We thereforecompared the response of mice stimulated with CEA and relatedproteins to those stimulated with LPS. In this study, systemic injectionof CEA stimulated the production of IL-6 in both athymic nude andBALB/c mice. The CEA treatment produced a moderate TNF-aresponse but did not generate a significant IL- la systemic response.We have shown previously that native CEA binds to Kupffer cells andalveolar macrophages in vitro and in vivo via a Mr 80,000 surfaceprotein receptor (23). After clearing CEA and removing the terminal

1::

% Maximum 60IL-6 Response 40

2:_.IL@AsIaIoCEA PELPK PELGK

a Groups of athymic nude mice and BALB/c were injected iv. with either 1 @sg/mouse

LPS; 0.05 ml of PBS; or 40 pg/mouse CEA, AAG, or asialoCEA in PBS. At 10 rein to24 h, blood was collected by cardiac puncture, and sera were assayed for IL-6, IL-la, andTNF-a by ELISA. Values are measured in pg/mI ±SE —,positive control for senim values(LPS). %, percentage of the maximal positive control response. ND, not determined.

100! .-100

%IL-6MaximalResponse

(@)

80

60

40

20

0

‘20

CEA (@tg/mouse)

-@--@-‘--v--

Fig. 2. Relative IL-6 production in athymic nude mice in response to an iv. injectionof CEA (0—40 @sg/mouse)in LPS-free PBS. Four- to 5-week-old athymic nude mice wereinjected iv. through the dorsal tail vein, and blood was collected by cardiac puncture atthe times indicated. Serum assayed for IL-6 concentrations by ELISA. Values wereplotted after deducting the untreated control values. Data are means; bars, SE. Results arecompared to the incidence of mice with liver colonies after a pretreatment with CEA,followed 30 mm later by an intrasplenic injection of 5 X l0@KM-l2c human CRC cells.The percentage of mice with liver colonies was determined at 6—8weeks after implantation and confirmed histologically.

1 @tgof LPS induced an IL-la response, which peaked at 404 ±6pg/ml within 1 h (Table 1). CEA induced a 156 ±42 pg/ml peak inTNF-a at 1—2h, which was only 3% of the peak produced in responseto LPS (4989 ±377 pg/ml at 1—2h, Table 1). AsialoCEA injectiondid not stimulate production of either IL-la or TNF-a. Thus, systemicCEA stimulated a strong IL-6 response but did not significantly stimulateeither TNF-a or IL-la production in athymic nude mice (Table 1).

The effect of CEA was also studied in BALB/c mice to confirm thatmature T cells did not affect the cytokine response pattern. CEA,AAG, and LPS injected i.v.-stimulated serum IL-6 responses thatwere similar to those seen with athymic nude mice (Table 1).

Effect of an Inhibitor of Tyrosine Phosphorylation on SystemicCytokine Production. Because LPS stimulates TNF-a and IL-6 production through tyrosine phosphorylation-dependent pathways (27,28), we assessed the effect of the inhibition of tyrosine kinase activityon the induction of IL-6 by CEA in Kupffer cells. Given that genisteinblocks LPS induction of the transcription of IL-lfl, IL-6 and TNF-ain both human and murine mononuclear phagocytes (26), we determined whether genistein affects the ability of CEA to induce IL-6 andTNF-a in mice. Pretreatment with genistein (100 mg/kg) diminished

Fig. 3. Specificity of IL-6 response. IL-6 production in athymic nude mice wasdetermined in response to an iv. injection of asialoCEA, PELPK-albumin, or PELGK (allat 40 @sg/mouse)in endotoxin-free PBS. Four- to 5-week-old athymic nude mice wereinjected iv. through the dorsal tail vein, and blood was collected by cardiac puncture. Serawere assayed for IL-6 concentrations by ELISA. Data are means; bars, SE.

4434

DISCUSSION

on April 17, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

U) Z < < < Ca U)n w ‘U ‘U ‘U 0. a.a. C, 0@ 0 @j -I

.2@ +

.5 z z. ‘U ‘U4 C, C,

CARCINOEMBRYONICANTIGEN INDUCTION OF CYTOKINES

A

IL-6(pg/mL)

B

TNF-a(pg/mL)

release IL-6, IL-la, and TNF-a in response to CEA and PELPKalbumin in vitro. The lack of an observed systemic IL-la effect doesnot rule out the possibility that IL-la is produced within the liver butis absorbed through local autocrine or paracrine mechanisms withinthe liver and does not reach the systemic circulation. Similarly, theabsolute value of the systemic TNF-a response may be underrepresented due to the presence of soluble TNF-a receptors, which reducethe amount of circulating TNF-a (29).

We then sought to begin to characterize the pathway through whichCEA induces the cytokine response after binding to Kupffer cells.LPS induces TNF-a, IL-6, and IL-1j3 in macrophages and monocytesafter binding to CD14 in concert with LPS-binding protein (31).Furthermore, the LPS time to peak response of cytokine production is

usually 2 h or longer (Table 1). In contrast, the induction of IL-6,TNF-a, and IL-la after i.v. CEA injection was only 1 h (Table 1).LPS induces tyrosine phosphorylation of a number of proteins, including the src family of tyrosine kinases (32). Others have shownthat tyrosine kinase inhibitors such as genistein and herbimycin Adecrease the amount of TNF-a, IL-6, and IL-la produced (26, 33). Inour study, systemic pretreatment with genistein decreased the IL-6response to CEA by 46% and to LPS by 47%. This suggests that theability of both CEA and LPS to stimulate the production of IL-6involves tyrosine phosphorylation. In contrast, when TNF-a wasmeasured in genistein-pretreated mice, the response to LPS wasdecreased markedly, whereas the response to CEA was increased3-fold. Thus, the pathways by which CEA induces production ofTNF-a may be different from that used by LPS.

The phosphorylation pattern of Kupffer cells exposed to CEA invitro further illustrates the difference between activation with LPS andCEA. A Mr 80,000 protein was phosphorylated on tyrosine, whereasa Mr 60,000 protein appeared to be dephosphorylated. These proteinshave not yet been identified. However, the Mr 80,000 protein may bethe CEA-binding protein, although it is not known whether this

220

97

5000

4000

3000

2000

1000

U) Z 4 4 .4 0 CaCa Ui Ui @j LU@0. CS C@@ 0 @j -J

+ +:@. z z4 Lu ‘U

C, C,

Fig. 4. A, effect of genistein on the induction of IL-6. Athymic nude mice werepretreated with 100 mg/kg genistein by i.p. injection for 1 h. The mice were then injectedwith 0.1 ml of PBS, CEA (40 gig/mouse), or LPS (1 pg/mouse). After I h, blood wascollected by cardiac puncture, and sera were assayed for IL-6 concentrations by ELISA.GEN, genistem. Data are means; bars, SE. B, effect of tyrosine kinase inhibition on theinduction ofTNF-a. Athymic nude mice were pretreated with 100 mg/kg genistein by i.p.injection for 1 h. The mice were then injected with 0.1 ml of PBS, CEA (40 pg/mouse),or LPS (1 @.sg/mouse).After I h, blood was collected by cardiac puncture, and sera wereassayed for TNF-a concentrations by ELISA. Values were plotted after deducting theuntreated control values. GEN, genistein. Data are means; bars, SE.

sialic acids from its oligosaccharide side chains, Kupffer cells releaseasialoCEA. The removal of the terminal sialic acid residues allows theterminal galactose residues of asialoCEA to be exposed to the hepatocyte asialoglycoprotein receptors (24). Because Kupifer cells bindthe amino acid sequence PELPK on CEA, asialoCEA will also bind toKupffer cells in vitro. However, because the affinity of hepatocyteasialoglycoprotein receptors is greater for asialoCEA than nativeCEA, hepatocytes will preferentially bind asialoCEA in vivo (24). Theinability to produce significant amounts of any of the three cytokinesafter asialoCEA injection suggests that Kupffer cells rather thanhepatocytes are the essential mediators of the ability of CEA to inducea cytokine response. The significant IL-6 response after PELPKalbumin injection but not after PELGK-albumin further supports thisconcept, because PELPK does not bind to other cells. These findingscomplement our previous report (22) that human and rat Kupffer cells

C0

NTR L C0 P EL S A

66@

46

30

Fig. 5. Effect of CEA on tyrosine phosphorylation; tyrosine phosphorylation pattemsin 2 X l0@isolated murine Kupffer cells treated with LPS (1 @sg/ml)or CEA (40 @.sg/ml).Western blots with monoclonal antibody to phosphotyrosine were performed as describedin “Materialsand Methods―on lysates of Kupffer cells following a I-h treatment in vitrowith LPS or CEA.

4435

on April 17, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

CARCINOEMBRYON1CANTIGEN INDUCTION OF CYTOKINES

protein can be phosphorylated (21). A number of signal transductionproteins in the Mr 60'000 range have been described, most notably thesrc family of protein tyrosine kinases (32, 34). Additional work isneeded to better define the nature of these proteins.

The functional role of IL-6 in the metastatic process has not yetbeen defined clearly. It was described originally as the hepatocytestimulating factor and may be an autocrine growth factor for melanoma (35), multiple myeloma, renal cell carcinoma (36), and cholangiocarcinoma (37). Furthermore, IL-6 is elevated in the serum of

many patients bearing tumors in proportion to tumor burden (38). Ina rat carcinoma model, IL-6 production was associated with metastaticpotential (17, 39). Because IL-6 with TNF-a and IL-la may increasethe expression of intercellular adhesion molecule 1 and endothelialleukocyte adhesion molecule 1 on endothelial cells (18), IL-6 mayincrease the retention of tumor cells within the liver. IL-6 may alsoinhibit apoptosis in certain cell types (40, 41). Additional work needsto be done to define the role of IL-6 in the interaction between tumorcells and the hepatic sinusoidal endothelium.

In summary, CEA induces a significant dose-dependent lL-6 response.CEA also induces a TNF-a response in both athymic nude mice andimmunocompetent mice after binding to a Mr 80,000 CEA-bindingreceptor. The signal transduction events involve tyrosine phosphorylation

events that may be different from those induced by LPS. As a result, CEAin the serum may enhance the ability of weakly metastatic CRC cells tocolonize the liver through the induction of IL-6, which then inhibits the

ability of host to kill weakly metastatic tumor cells. As a result, IL-6 mayhave a pivotal role in the neoplastic progression of CRC.

REFERENCES

I. Gold,P., and Freeman,S. 0. Demonstrationof tumorspecificantigenin humancolonic carcinomata by immunological tolerance and absorption techniques. J. Exp.Med., 121: 439—446,1965.

2. Goslin, R., Steele, G. D., Maclntyre. J., Mayer, R., Sugarbaker, P., Cleghom, K.,Wilson, R., and Zamcheck, N. The use of preoperative plasma CEA levels for thestratification of patients after curative resection of colorectal cancers. Ann. Surg.,192: 747—751,1980.

3. Wanebo, H. J., Rao, B.. Pinsky, C. M., Hoffman, R. G., Steams, N., Schwartz, M. K.,and Oettgen,H. F. Preoperativecarcinoembryonicantigenlevel as a prognosticindicator in colorectal cancer. N. Engl. J. Med., 299: 448—451. 1978.

4. Macnd, C. G., Fleming, T. R., Macdonald, J. S., Haller, D. G., Laurie, J. A., andTangen, C. An evaluation of the carcinoembryonic antigen (CEA) test for monitoringpatients with resected colon cancer. 1. Am. Med. Assoc.. 270: 943—947,1993.

5. Thomas, P., Gangopadhyay, A., Steele, G. D.. Andrews, C., Nakazato, H., Oikawa, S.,and Jessup, J. M. The effect of transfection of the CPA gene on the metastaiic behaviorof the human colorectal cancer cell line MIP-lOl. Cancer Lett., 92: 59—66,1995.

6. Hashino,1., Fukuda.Y., Oikawa,S., Nakazato,H., and Nakanishi,1. Metastaticpotential of human colorectal carcinoma SW1222 cells transfected with cDNAencoding carcinoembryonic antigen. Clin. Exp. Metastasis, 12: 324—328, 1994.

7. Thompson. J. A. Molecular cloning and expression of carcinoembryonic antigen genefamily members. Tumor Biol., 16: 10—16,1995.

8. Rutishauer, U., Achesion, A., Hall, H. K., Mann, D. M., and Sunshine, J. The neuralcell adhesion molecule (NCAM) as a regulator of cell-cell interactions. Science(Washington DC). 240: 53—57,1988.

9. Benchimol,S., Fuks,A.,Jothy,S., Beauchemin,N.,Shirota,K.,andStanners,C. P.Carcinoembryonic antigen, a human tumor marker. functions as an intercellularadhesion molecule. Cell, 57: 327—334,1989.

10. Hostetter,R. B.,Campbell.D. E.,Chi. K., Kerckhoff,S., Cleary,K. R., Ulrich,S.,Thomas, P., and Jessup, J. M. Carcinoembryonic antigen enhances metastatic potential of human colorectal carcinoma. Arch. Surg.. 125: 300—304, 1990.

I 1. Kuijpers, T. W.. Hoogerwerf, M., van der Laan, L., Nagel, G., and van der Schoot,E. CD66nonspecificcross-reactingantigensare involvedin neutrophiladherencetocytokine-activated endothelial cells. J. Cell Biol., 118: 457—466,1992.

12. Ohannesian, D. W., Lotan, D., Thomas, P., Jessup, J. M., Fukuda, M., Gabius, H. 1.,and Lotan, R. Carcinoembryonic antigen and other glycoconjugates act as ligands forgalectin-3 in human colorectal carcinoma cells. Cancer Res., 55: 2191—2199,1995.

13. Jessup, J. M., Petrick, A. T., Toth, C., Ford, R., Meterissian, S., O'Hara, C. J., Steele,G. D., and Thomas, P. Carcinoembryonic antigen: enhancement of liver colonizationthrough retention of human colorectal carcinoma cells. Br. J. Cancer, 67: 464-470, 1993.

14. Hostetter, R. B., Augustus, L. B., Mankarious, R., Clii, K., Fan, D., Toth, C., Thomas,P., and Jessup. J. M. Carcinoembryonic antigen as a selective enhancer of colorectalcancer metastases. J. NatI. Cancer Inst., 82: 380—385, 1990.

15. Mannel, D. N., Orosz, P., Hafner, M., and Falk, W. Mechanisms involved inmetastasis enhanced by inflammatory mediators. Circ. Shock, 44: 9—13,1994.

16. Obdenakker, G., and Van Damme, J. Cytokines and proteases in invasive processes:molecular similarities between inflammation and cancer. Cytokine, 4: 251—258,1992.

17. Takeda, K., Fujii. N., NiUa, Y., Sakihara, H., Nakayan. K., Rikiishi, H., and Kumagai,K. Murine tumor cells metastasizing selectively in the liver: ability to producehepatocyte-activating cytokines interleukin-l and/or -6. Jpn. J. Cancer Res., 82:1299—1308,1991.

18. Ueda, T., Shimada, E., and Urakawa, T. Serum levels of cytokines in patients withcolorectal cancer: possible involvement of interleukin-6 and interleukin-8 in hematogenous metastasis. J. Gastroenterol., 29: 423—429, 1994.

19. Fearson, K. C. H., McMillan, D. C., Preston, T., Winstanley, F. P., Cruickshank,A. M., and Shenkin, A. Elevated circulation interleukin-6 is associated with anacute-phase response but reduced fixed hepatic protein synthesis in patients withcancer. Ann. Surg., 213: 26—31,1991.

20. Gangopadhyay, A.. Bazhenova, 0., Kelly, T., and Thomas, P. A. Carcinoembryonicantigen induces cytokine expression in Kupffer cells: implications for hepatic metastasis from colorectal cancer. Cancer Res., 56: 4805—4810,1996.

21. Thomas. P., Petrick, A. T., Toth, C. A., Fox, E. S., Elting, J. J., and Steele. G., Jr. Apeptide sequence on carcinoembryonic antigen binds to a 80 kD protein on Kupifercells. Biochem. Biophys. Res. Commun., 188: 671—677, 1992.

22. Gangopadhyay, A., and Thomas, P. Receptor mediated endocytosis of carcinoembiyonic antigen by Kupifer cells: recognition of a penta-peptide sequence. Arch. Biochem. Biophys., 334: 151—157,1996.

23. Thomas, P., and Toth, C. A. Carcinoembryonic antigen binding to Kupffer cells is viaa peptide located at the junction of the N-terminal and first loop domains. Biochem.Biophys. Res. Commun., 170: 391—396.1990.

24. Thomas, P., and Hems, D. A. The hepatic clearance of circulating native and asialocarcinoembryonic antigen by the rat. Biochem. Biophys. Res. Commun.. 67: 1205—1209, 1975.

25. Whitehead, P. H., and Sammons, H. G. A simple technique for the isolation oforosomucod from normal and pathological sera. Biochem. Biophys. Acta., 124:209—211,1966.

26. Geng Y., Zhang. B., and Lotte, M. Protein tyrosine kinase activation is required forlipopolysaccharide induction of cytokines in human blood monocytes. J. Immunol.,151: 6692—6700, 1993.

27. Weinstein, S. L., Gold, M. R., and De Franco, A. L. Bacterial lipopolysaccharidestimulates protein tyrosine phosphorylation in macrophages. Proc. NatI. Acad. Sci.USA. 88: 41—48,1991.

28. Weinstein, S. L., Sangher, J. S., Lemke, K., Dc Franco, A. L., and Pelech, S. L.Bacterial lipopolysaccharide induces tyrosine phosphorylation and activation ofmitogen-activated protein kinases in macrophages. J. Biol. Chem., 267: 14955—14962, 1992.

29. Jessup, J. M., Kim, J. C., Thomas, P., Ishui, S., Ford, R., Shively, J. E.. Durbin, H.,Stanners, C. P., Fuks, A., Thou, H., Hansen, H. J., Goldenberg, D. M., and Steele, G.,Jr. Adhesion to carcinoembryonic antigen by human colorectal carcinoma cellsinvolves at least two epitopes. mt. i. Cancer, 55: 262—268,1993.

30. Wemer-Wasik, M., von Muenchhausen, W., Nolan, J. P., and Cohen, S. A. Endogenous interferon a/@3produced by murine Kupffer cells augments liver associatednatural killing activity. Cancer Immunol. Immunother., 28: 107—115, 1989.

31. Wright, S. D., Ramos, R. A., Tobias, P. S., Ulevitch, R. J., and Mathison, J. C. CD14,a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein.Science (Washington DC), 249: 1431—1433,1990.

32. Beaty. C. D., Franklin, 1. L., Uehara, Y., and Wilson, C. B. Lipopolysaccharideinduced cytokine production in human monocytes: role of tyrosine phosphorylation intransmembrane signal transduction. Eur. J. Immunol.. 24: 1278—1284.1994.

33. Matsuyama, S., Koide, Y., and Yoshida, T. 0. HLA class II molecule-mediated signaltransduction mechanism responsible for expression of IL-1f3 and TNF-a genes inducedby a staphylococcalsuperantigen.Eur.J. Immunol.,23: 3194—3202,1993.

34. Hatakeyama, S., Iwabuchi, K., Ogasawara, K., Good, R. A., and Onoe, K. The murinec-fgr gene product associated with Ly6C and p70 integral membrane protein isexpressed in cells ofa monocyte/macrophage lineage. Proc. NatI. Acad. Sci. USA, 91:3458—3462,1994.

35. Lu, C., Sheehan, C.. Rak, J. W., Chambers, C. A., Hozumi, N., and Kerbel, R. S.Endogenous interleukin 6 can function as an in viva growth-stimulatory factor foradvanced-stage human melanoma cells. Clin. Cancer Res., 2: 1417—1425,1996.

36. Blay, J-Y., Mercatello, A., Negrier, S., Philip, T., and Favrot, M. C. Serum levels ofIL-6 receptor and IL-6 as independent prognostic factors for survival in metastaticrenal cell carcinoma. Proc. Annu. Meet. Am. Assoc. Cancer Res., 36: 21 1, 1995.

37. Carty, S. E., Frezza, E. E., Selby, R. R., Booth, A., Brumfield, A. M., and Lotze, M. T.Marked elevation of serum IL-6 levels in patients with cholangiocarcinoma. Proc.Annu. Meet. Am. Assoc. Cancer Res., 35: 17, 1994.

38. Mouawad, R., Benhammouda, A., Rixe, 0., Antoine, E. C., Borel, C.., Weil, M.,Khayat. D., and Soubrane. C. Endogenous interleukin 6 levels in patients withmetastatic malignant melanoma: correlation with tumor burden. Clin. Cancer Res., 2:145—149,1996.

39. Reichner, J. S., Mulligan, J. A., Palla, M. E., Hixson, D. C., Albina, J. E., and Bland,K. I. Interleukin-6 production by rat hepatocellular carcinoma cells is associated withmetastatic potential but not with tumorigenicity. Arch. Surg., 131: 360—365,1996.

40. Chauhan, D., Kharbanda, S., Ogata, A., Urashima, M., Teoh, G., Robertson, M.,Kufe, D. W., and Anderson, K. C. Interleukin-6 inhibits Fas-induced apoptosis andstress-activated protein kinase activation in multiple myeloma cells. Blood, 89:227—234,1997.

41 . Reittie, J. E., Yong, K. L., Panayiotidis, P., and Hoffbrand, A. V. Interleukin-6inhibits apoptosis and tumour necrosis factor induced proliferation of B-chroniclymphocytic leukaemia. Leuk. Lymphoma, 22: 83—90,1996.

4436

on April 17, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1997;57:4432-4436. Cancer Res   Kirsten H. Edmiston, Aniruddha Gangopadhyay, Yutaka Shoji, et al.   Carcinoembryonic Antigen

Induction of Murine Cytokine Production byIn Vivo

  Updated version

  http://cancerres.aacrjournals.org/content/57/19/4432

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/57/19/4432To request permission to re-use all or part of this article, use this link

on April 17, 2021. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from