UNIFORMED SERVICES UNIVERSITY OF THE HEALTH SCIENCES 4301 JONES BRIDGE ROAD

GlADUAT&EDVCATlOI'I (lOIlM3913

FAX: {JOl):m.4m

BETHESDA, MARYLAND 20814-4799

APPROVAL SHEET

Title of Dissertation: "Identification of the First Receptor for a Pregnancy Specific Glycoprotein. Tetraspanins find their Ligand"

Name of Candidate: Roseann Waterhouse Doctor of Philosophy Degree 13 March 200 1

Dissertation and Abstract Approved:

Gabrie a Dv ler, Ph.D. Department of Pathology Committee Member

George Cox, Ph.D. Department of Phannacology Committee Member

~D. National Institutes of Health

c~¥r / Phi Ip Grimley, .D., Department of athology Commiuee Member

Printed on G> Re~led Piper

Date

Date

Date

ttl ate

Date

The author hereby certifi es that the use of any copyrighted materia l in the di ssertation enti tled:

" Identifi cation of the First Receptor for a Pregnancy Specific Glycoprotein Tetraspanins find their Ligand"

beyond brief excerpts is with the permission of the copyright owner, and wi ll save and hold hannlcss the Uniformed Services University of the Health Sciences from any damage which may ari se from such copyright violations.

Roseann Marie Waterhouse Department of Pathology Unifonned Services University of the Health Sciences

ABSTRACT

Identification of the First Receptor for a Pregnancy Specific Glycoprotein. Tetraspanins find their Ligand.

Roseann Maric Waterhouse, Doctor of Philosophy, 200 1

Thesis Directed by Gabriela Dvekslcr, Associate Professor, Department of Pathology

Pregnancy specific g lycoprotc ins (PSGs) arc a family of secreted proteins produced by

the placenta, which have been shown to be essen ti al for pregnancy success. The ability

of human PSGs to induce anti·inllamll1atory cytokines has led to the hypothesis that these

proteins may function to protect the fctus from attack by the maternal immune system.

With the purpose of developing an animal model to study the function of rSGs, we have

studied the effects of murine PSG 17 0 11 macrophages and wc have cloned its receptor.

RA W 264.7 ce ll s and peritoneal macrophages were treated with recombinant PSG 17N.

which consists of the N-domain of PSG 17. PSG 17N induced production of IL-l 0 and

IL-6 at the protein and RNA levels in these cel ls. Secretion of TGFJ31 and PGE2 was also

induced upon treatment with PSG 17N. We then examined the PSG 17-RA W ee l! surface

binding interaction. Scatchard anal ys is revealed that there are approximately 1770

binding sites per cell with a Kd of2.2 X 10.11 M. For the purpose of cloning the PSG 17

receptor, we screened a RA W cell eDNA library by panning. The receptor was identified

as CD9, a member of the tetraspanin superfami ly. The speciticity of the C09-PSG 17

interaction was confirmed by ELISA and fl ow cytometry in C09-transfected cells.

Furthennore, binding of PSG 17 to C09-expressing ce ll s was blocked with anti-C09

antibodies. We then tested whether murine PSG 18 and 19, which share with PSG 17 the

ability to induce cytokines in macrophages. use CD9 as their receptor. These proteins do

II

not bind 10 CD9. In conclusion, we have identified the first receptor for a PSG as well as

the firsl natural ligand for a member of the tetraspan in superfamily.

1\1

IDENTIFICATION OF THE FIRST RECEPTOR

FOR A PREGNANCY SPECIFIC GLYCOPROTEIN.

TETRAS PAN INS FIND THEIR LIGAND

By

Roseann Marie Waterhouse

Thes is/dissertation submitted to the Faculty of the Department of Pathology of the

Unifomlcd Services University of the Health Sciences in partial fulfillment of the requirements for the degree of

Doctor of Philosophy 200 1

IV

ACKNOWLEDGEMENTS AND DEDICATION

I am very gratefu l to Sara Snyder. Jennifer Wesse ll s, David Wess ner, and Kimberley White for all of thei r adv ice, encouragement, and technical ass istance in this research. I would also like to thank my advisor, Dr. Gabriela Dveksler. for her support, guidance and tolerance through the times of hope and frustration, wh ich accompanied thi s resea rch.

This dissertation is dedicated to my parents. George and Patricia Parse li s. and to my husband. Scott Waterhouse, for their suppOI1, encouragement and fa ith, throughout my years of work toward thi s degree.

v

l. Introduction Significance

Table of Contents

The PSG Family is a Subfamily in the Carcinoembryonic Antigen Family

Gene and Protein Structure for I-iuman CEA family members

Expression of Human Pregnancy Specific Glycoprorcins I. Sites of expression 2. Human PSG Structure and function

Murine Pregnancy Speci fi c Glycoproteins I. Sites of expression 2. Structure and fun ct ion

A Murine Model for PSG Function

Pregnancy Specific Glycoprotcins in Other Animal Species

Placental PSGs: Expression of human PSGs in nonnal pregnancy and in cancer

A Brief Overview of the Multifaceted Phenomenon of Fetal Tolerance

Page

2

4

8

\0

II

12

during Pregnancy 14

Anti-inflammatory Thl-Type Immune Response During Pregnancy as a Tolerance Mechanism 15

Other Potential Mechanisms of Pregnancy Maintenance 21

CD9 is a member of the Tetraspanin Superfamily 22

Placental Implantation, Tctraspanins and Pregnancy 25

Specific Aims, Significance, Hypothes is and Approach 29

Results 30

II. Paper #1: Pregnancy Specific Glycoprotein 17 Binds with High Affinity to Murine Macrophages and Induces Production of Cytokines in vitro. 32

Introduction 33

VI

(Paper # 1 Continued) Materia ls and Methods 35

Resu lts

Allimals and Cell Cllltllre Reagents SlIhcloning oIPSG 17FL iI/to Alkaline PllOspJwtase (AP)- Tag4 AP-Fusioll Proteill Binding Assays Alla~vsis o/Cytokine indllclion by PSG-treated macrophages and

RA W 264. 7 cells SelJli-qlfallfitative RT-PCR Data Allall'sis

PSG 17 induces secretion of1L-1 () ill I11urine lI1(JcropJwges in a dose dependent mal1l1er

39

PSGs illdllce production C?f IL-6 alld TGF/JI in murine lIIacropJwges PSG illduction oflL-6 and IL-I{) requires de 1I0VO protein synthesis PSG 17 illduces the eicosanoid. Prostaglandin E! ill RA W 264. 7 cells PSG 17 binds with high qIJil1ity to a limited number oIbilie/ing sites 011

RAW 264.7 cells

Discussion 48

References 54

III. Paper #2: Murine CD9 is the Cellular Receptor for Pregnancy Specific Glycoprotein 17 56

Introduction 57

Materi als and Methods 59

Results

Reagents Cell Cullllre Generation q{a murine macrophage eDNA Jibraty RecovelY 0/ cDNA clones by Panning Deteclion q{ PSG 17 Binding to tran.~{ec{ed HEK 293T ce1ls

by ELISA In Silll Roselling Assay Flow cYlOmel1y Alkaline phosphatase binding assays

1dellt{fieQtioll q{CD9 as the cellular receptor/or PSG 17 Binding qlAlkaline phosphatase-PSG17 (AP-PSG17) to CD9 Murine PSG J 8 and Murine PSG 19 do 110t hilld to CD9

V II

64

(Paper #2 Continued) Discussion

References

IV. Discussion PSG Induction ofCytokines in Pregnancy

IL-In alld Macrophages IL-/o Expressioll in Trophoblast,\' ",terlel/kill 6 and Pregllall(v The Importance (?fTGF-pl during Pregnancy Prostaglandin E: and Pregnancy

CD9- PSG 17 Interaction Sites o/Expressioll f?fCD9 CD9 and T cells CD9 Expressioll in Non-Hematopoietic Tissues

Signali ng Mechanisms The PSG 17N signaling cascade CD9 Signaling Events

Working toward a Comprehesive Mouse Model of PSG Function Generatioll % Murine Allti-PSG Alltibody Knockouts and Transgenes Identijving additiollal humall alld muril/e PSG receptors

V. Appendix I PSG Nomenclature

VI. Appendix 2 Purification of PSG 17N-Myc His

VII . References for Thesis Introduction and Discussion

VIII

72

74

76

79

86

88

93

96

List of figures

Section Page Number

Iniroduci ion :

Figure 1. Schematic representation of four members of the TelnlSpanin Superfamily (TM4SF)

Figure 2. Schemati c Diagram of the Human Placenta

Paper #1: Pregna ncy Specific Glycoprotein 17 Binds with High Affinity 10 Murine Macrophages and Induces Production of Cylokines ill vitI"(}.

Figure I . PSG 17N induces IL-I 0 secretion in murine macrophages

23

27

and RAW 264.7 cells 4 1

Figure 2. PSG I7N induces IL-6 secretion in murine macrophages and RA W 264.7 cells 42

Figure 3. PSG 17N induces TGF~ ! sec retion by murine macrophage and RAW 264.7 cells 44

Figure 4. Cycloheximide inhibits PSG 17N induct ion of lL-6 and lL- IO 45

Figurc 5. PSG 17N Induced PGE, in RA W 264.7 cells

Figure 6. Scatchard Analysis of PSG 17 Binding to RA W 264.7 cells

Paper #2: Murine C D9 is the Cellular Receptor for Pregnancy Specific C lycoprotein 17

Figure I . Binding of PSG 17N to murine CD9 transfected 293T cell s

Figure 2. Anti-CD9 antibodies inhibi t PSG 17 binding

Table I . T reatments used in Flow Cytometric Analysis of PSG binding to CD9 in HEK 293T

Figurc 3. FACS analys is of PSG 17N-Myc-His bindi ng to CD9 transfected 293T cel ls

Table 2. Treatment protocol for PSG binding to CD9 in BI-IK-2 1 cells by flow cytometry

IX

47

65

65

68

68

69

(Figures in Paper #2 cont inued)

Figure 4 . FACS analysis of PSG 17N binding to BHK-2 1 cell s expressing murine C0 9

Figure 5. Full length PSG 17 binds to C0 9 transfectcd cel ls

Figure 6 . PSG 18N and PSG 19N do not bind to C0 9

x

69

71

71

I. Introduction

Significance

Pregnancy specific glycoprotcins (PSGs) arc a famil y of secrctcd proteins

produced by the placenta. They were di scovered in 197 1 in the blood of pregnant women

[1]. PSGs (lrc first detected 3 to 4 days post fertilization. coinciding with implantation of

the blastocyst in lO the uterine wa ll [2]. Matemal serum concentration of these proteins

increases exponentially during pregnancy reaching levels of 200-400 IJglml at term [3, 4],

making PSGs the most prevalent "pregnancy- related" protein in matemal semrn.

Reduced express ion of these proteins correlates with fetal hypoxia, fetal growth

retardation. pre-eclampsia. and spontaneous abortion f5-8]. Despite their identification

thirty yea rs ago, in vitro biological functions for PSGs have only been identified in the

past few years [9-12], and their phys iological role in pregnancy has yet 10 be defined.

The research reported in these manuscripts contributes to the accumulating knowledge

regarding PSG immuno-regulatory function ill vitro. and, importantly, th is is the first

report to identify a receptor for (I PSG.

The PSG Family is a Subfamily in the Carcinoembryonic Antigen Family

Screening of human placental eDNA libraries revealed that the pregnancy

specific glycoprotein genes are highl y homologous [1 3-15]. The PSG gene family is

class ified by sequence homology as a subfamily in the Carcinoembryonic antigen (CEA)

gene family. The CEA gene family belongs to Immunoglobulin (Ig) superfamily [16].

Transcripts of CEA gene famil y members code for proteins containing repeating domains

each consisting of ~-shect stmcture and the conserved di sulfide bridges of an 19 fold, all

characteristics of th is superfamily. The CEA gene fami ly can be subdivided into the

CEAI nonspecific cross~ rcact i ng antigen (NCA) subfamily and the PSG subfamil y. The

PSG subfami ly only contains pregnancy spec ific glycoprole ins. while the CEA subfamily

is comprised of carc inoembryonic an tigens (CEA). bi liary glycoproteins (BGP).

nonspecific cross~reacting ant igen (NeA) and six OI her CEA fami ly members (CGMs)

(reviewed in [17 ~ 19]). All 29 members of the human CEA gene famil y are localized to

chromosome 19 [20, 2 1]. There are 12 CEAINCA genes. II PSG genes {22] and 6

CGMs (CGM 1 3 ~ 18) [23]. Four of the 12 CEA genes arc pseudogcncs while none or lhe

CGM genes have shown to be expressed. Importantly, there are no known pseudogenes

in the human PSG fam il y and expression of all I I PSG genes has been reported [23, 24].

Gene and Protein Structure for Human CEA family members

The members of the CEA gene family have similar gene organizat ion. The first

exon (designated UN) codes for the S'UTR and part orlhe leader signa l pept ide (2 1 aa).

Exon two encodes the remainder of the signal peptide and the first domain of the mature

protein. lemled the "N~domai n" ( 107-1 10 aa). This domain is homologous to

Immunoglobulin variab le (IgV) domains. Additiona l domains. which are I g~constant

(JgC) -like, arc encoded by single exons. These IgC- li ke domains are designated "A" (92

aa) and " B" (86 aa). The number of IgC-like domains is gene-specific, ranging from zero

to six. Each IgC- li ke domain contains at leaslIwo cysteine residues, which are assumed

to stabili ze the immunoglobulin-like fold by forming a disulfide bridge. Significant

nucleotide and amino ac id conservation among the internal repeat domains of PSG and

2

CEA genes suggests that both families evolved recently by the dupli cat ion ora single

primordial gene [25].

The CEA/PSG subfamily division also reflects some differences in mature protein

structure. The proteins from both subfa milies are structurall y s imilar, containing mult ip le

Ig-like domains and are highl y glycosylaled, however, CEA/NCA proteins are attached 10

the plasma membrane by a hydrophobic transmembrane region at the C' tenninus, or

through a covalen t linkage to membrane-bound glycosylphosphat idyl- inos itol [26].

However, in PSGs, a truncated , largely hydrophili c region replaces thi s C-temlinal

domain , dictating sec retion of the proteins by the ce ll [1 8]. Slight differences in post­

translational modi fica tion ofCEA family members further indicate separation between

the two subfamilies [25). At least 30% of the molecular weight ofPSGs can be attributcd

to the addition of carbohydrate residues, whereas the total glyeosylation of other CEA

famil y members can range from 30-60% [ 17]. In contrast to the differences between

subfamilies, the N-domain ofCEA gene family proteins are highl y conserved within each

subfamily, exhibiting over 90% nucleotide sequence similarity [27]. (N-domain

alignments and diagrams of human, mouse, rat and guinea pig CEA and PSG subfamily

structura l domains can be viewed at the CEA homepage www.ukl. un i­

freiburg.delimrnz/cca/ .) Nucleotide sequence homology between the two CEA

subfamilies is less conserved in the N domain (65- 75%). The N-domains of PSG, NCA,

and CEA proteins mediate binding to their putative li gands, which implies thi s domain is

critica l for biological function [ II , 12, 28, 29].

3

Expression of Human Pregnancy Specific Glycoproteins

I. Sites of express ion

Human PSGs are primarily expressed by syncytiotrophoblasts [30] and PSG

transcripts have also been identified in amniollic and chorionic tissue [3 1]. which

suggests that these proteins playa key ro le in pregnancy. Although the most abundant

source of PSGs is the placenta [32J. expression of PSG transcript s is not limited to

pregnancy [33]. Expression of hu man PSG transcripts has also been described in

intestine [34], testi s [35], uterus [36], salivary gland [37]. endometrial ti ssue [10], bone

marrow plasma [38], and breast tumors [39, 40] . The presence of PSGs has been

reported in speci fi c ccJl types including fibroblasts in culture [41. 42], myelo id cell lines

[43 ,44], on the surface and in cytoplasm of granulocytes [45]. and in cells of the

myelomonocyti c lineage [38]. PSGs arc produced by trophoblastic tu mors and non­

trophoblastic tumors such as choriocarcinoma and breast canccr [39. 46, 47]. PSG

transcripts havc been cloned from first trimester fe tal li ver, howevcr PSG mRNA has not

been iso lated from adult li ver [48]. The function ofPSGs outside of pregnancy is

unknown . Further analysis of PSG function and receptor location may bring to light the

significance of these external sites of expression.

With primary sequence conservation between all members of the CEA family, the

accuracy of some immunological methods used to identify expression si tes is

questionable. Experiments using anti- sera including immunoprecipitation,

immunohistochemical sta ining and Western blotting may be fl awed as antibodies to CEA

fam ily proteins have been shown to cross-react between family and subfamily protein

4

members [ 18, 49]. In add ition , o ligonucleotide probes have been shown to cross-react

with CEA and NCA famil y members [50].

2. Human PSG Structure and function

Each human PSG transcript encodes a 34 ami no acid (aa) leader (L) that is

cleaved upon secretion of the protein, an N-tenninallgV- likc domain (108- 109 aa )

designated "N", two or three intemallgC- like domains [A lIA2 (93aa) and/or 81 / 82

(85aa)J and a C-temlinal domain (C) (2 - 8 1 aa) [20]. PSG genes are alternative ly sp li ced

leading to variation in domain structure lor individual IlSGs. PSG mR NA organizat ion is

primarily either L-N-A i-A2-B2-C, L-N-A i-B2-C, or L-N-A2-B2-C [23]. All PSG

polypeptides have 324 to 435 amino acid backbones with multiple potential N-linked

glycosylation sites. Although splice variants ex ist, only three distinct serum PSG

fractions of 72, 64, 58 kilodaltons have been identifi ed by gel electrophoresis [ 15].

1·luman PSGs contain only onc Ig-variable-like domain, the N-domain, while the A and B

domains are all IgC-like. Exclud ing the C-tcmlinus, the nucleot ide sequences of human

PSGs are 90% homologous [20].

As is characteristic of the CEA famil y members, the PSG N-domain was

suspected to be responsib le for binding of these proteins to their ligands. The N-terminal

region of recombinant PSG6 has been shown to be suffi cient for binding to human

monocytes [ 12J. In addition, a 20 amino acid peptide deri ved from thi s domain binds to

promonocytic cc ll s ill vitro [51] suggesting that thi s domain is essen tial for PSG function .

The arginine- glycine-aspartic ac id (RGD) sequence motifi s found at a conserved

position within the N-domain of PSG 2, 3, 5, 6, 7, 9 and II . The significance of thi s

5

motif remains controversia l. RG D motifs are present in several proteins wi th integrin

ligands such as fi brinogen, vi troncctin , and fib ronectin [52]. Bascd on the presence of

thi s motif, proposed bio logical acti vit ies for PSGs have included ce ll -mediated binding

with ext racellu la r matrix [1 3,53). growth fac tor inh ibit ion [25. 54]. and inhibi tion of ce ll

migration (53] . To date none of these func tions has been demonstrated for PSGs in vivo

or in vitro.

However there is evidence in vitro to support an immuno-regulatory fun ction for

PSGs. In one stud y, PSG II which conta ins an RG D moti f. but not PSG I (which does

not contai n an RGD sequence). increased IL-IO production by LPS treated pri mary

monocytes and the promonocytic ce ll lines. U937 and THP- I, leading one research group

to suggest that the presence of the RGD motif in PSG I I was essential fo r binding and

biologicaf function [5 1]. In contrast, Ollr labo ratory has demonstrated that both PSG 11

and PSG I have similar biological func tions. Both of these proteins as well as PSG6

induced production of IL-6, IL- JO and TGF PI by human e lutriated monocytes ill vitro,

albeit wi th different kinet ics [12]. The disparity between our results and those of

Rutherford can poten tially bc exp lained by the quantity of protein used in each study.

Our results indicate that higher concentrations of PSG I and PSG II were necessary to

induce cytok ines production in these cells . The high levels of PSG expression during

pregnancy, together with their influence on monocyte cytokine production are consistent

with the proposed biological fun ction of fe tal protection from maternal immune rejection

[55,56].

PSGs have a lso been suggested to have a role in embryogenesi s. Murine

embryos co-cultured with recombinant human PSG I developed rapidly into morulae and

6

blastocysts [57). However, incubat ion with recombinant PSG3 did not increase ce llular

proliferation. These results suggest that PSG I and PSG3 play different roles in fetal

developmem, and that some human PSGs may cross-react with murine ligands. Further

evidence of human-murine cross-reactivity has been obtained from experiments where

recombinant human PSGs induced fl - l O. IL-6 and TGF~I in a mouse macrophage cell

li ne (RA W 264.7)[ 12]. Because differenlial tempora l expression of PSG proteins during

human pregnancy cannot be eas il y detenllined, a proliferative function for PSG I ea rl y in

gestation mayor may not be possible in human fetal development.

Outs ide the context of pregnancy. low basel ine PSG express ion levels of less than

0.5 J.lg/L have been measured in the serum ofmcn and non-pregnant women [32, 58].

The presence of PSG transcripts in fetal liver and adult bone marrow is suggesti ve of a

role for these proteins in hematogenesis. The developing liver is the primary site of fetal

hematopoiesis during weeks 10-3 1 of deve lopment, while bone marrow is the primary

location ofadu!! hematopoiesis. Feta l li ver and adult bone marrow PSG expression

implies potential involvement of these proteins in hematopoietic differentiation or ce ll

growth. In a stud y aimed to deten11 ine the ill vivo hcmatopoietic effects of PSGs. bonc

marrow was transplanted into irrad iated mice and then purified human PSGs were

injected at various doses fo r 20 da ys [9). The mice were bled every 2 days and ce ll

counts revealed that the groups treated with 13.9 or 139 11M had an increased platelet

count and a quicker white cell recovery rate compared to mice treated with control

protein. Unfortunately, a mixture ofPSGs was used in these experiments, making it is

impossible to detennine whether thi s effect was due to a single PSG protein or a

7

combinat ion ofPSGs. However, the results suggest that placental PSGs enhance platelet

and white blood ce ll counts supporting a ro le in hematogenes is.

With a vari ety of express ion sites and potential functions, human PSGs may bind

to one or more putati ve receptors, however a speci fi c molecular PSG binding partner has

not been identi fied. A peptide derived from PSG II , containing the RGD motif, has been

shown to bind to cells of the myelo id lineage, hut not to B or T ce ll li nes [5 1]. Human

monocytes treated with PSG I, PSG6 and PSG II induced a biological response, the

production of Th2 cylokincs [12]. Thus monocyte/macrophage cells express one or more

receptors fo r human PSGs. Crosslinking and immunoprecipitation experiments to

characterize the putati ve receptor on promonocyt ic TI-IP- I cells revealed a binding

product of 46 kilodaltons [51 ]. Whether this 46 kDa protein is a receptor for human

PSGs and is sufficient to trigger signaling mechanisms necessary for cytokine induction

or other putat ive PSG functions rema ins unknown.

Murine Pregnancy Specific Glycoproteins

1. Sites of expression

Murine PSGs are expressed by primary and secondary trophoblast giant cells of

mouse placenta earl y in gestation and then later by spongiotrophoblast ce lls and their

precursors. All murine PSGs are co-expressed by placental tissue 10- 17 days post

implantation [59], but expression at earl ier limes post-fertili zation can not bc ruled out.

RT-PCR analysis of adult murine tissues including kidney, lung, testi s, ovary, liver,

brain, thymus, heart and spleen did not ident ify transcripts for murine PSGI7- l9, 21 , 23

8

and 28 [ II " 60, 61 ]. The concentration o r murine PSG proteins in serum rrom pregnant

mice is not known.

2. Structure and function

Screening or a Y AC library revealed that there are 14 murine PSG genes. The

genes have been ide11l ifi ed based on pan ia l genomic sequence or the domain. These

genes are mapped to chromosome 7 by fluorescent in-sit u hybridization and are

expressed cxclusively in the placenta [ II , 60. 6 1]. The only murine PSG pscudogenc is

Cea6. which correlates with a deletion or onc nucleot ide in the leader sequence. resulting

in a shin in the reading rrame [60] . Chromosome 7 is syntcnic to human chromosome 19

where hu man PSGs are located [62]. It is difficult to experi menta ll y di stinguish between

di ffe rent murine gene family members as many PSG genes and mRNAs have not been

completely sequenced (Dveksler & Zimmennan. persollal comllluniC(/fio ll) . However.

full -length eDNAs have been identified for PSG 17- 19 [60J. PSG23 and PSG28 [6 1 J. In

these genes, as is characteristic of human PSGs. each exon codes for an individual protein

domain. All murine PSGs characteri zed so far contain four Ig-like domains [60. 63].

Murine PSGs differ from human PSGs in domain structure. conta ining three Ig variable­

like regions and a single Ig-constan t-li ke domain. The amino acid homology of lhe

murine PSG N-domains varies between 54-93%. Similar to the human N-domain RGD

motif, a highly conserved glycine residue flanked by one of four amino acids (RlH-G­

ElK) is found at a conserved site in the muri ne N-tenninal domain [60). This moti f is

present in all murine PSGs with the exception of PSG 19, in which the tri-peptide

asparagine-g lycine- lysine (NGK) is found at thi s site.

9

Our laboratory was the first to report a bio logical function for a murine PSG [ II J.

In an ill vitro study, murine macrophages were treated with a recombinant GST-PSG 18

fusion protein, produced using a baculovirus ex pression system. Ana lysis of levels of

RNA for the TH2 cytok ine Intcrleukin- IO was perfomled. Treatment wi tb both PSG I 8

and a truncated fonn of the prote in consisting of only the N-domain (PSG I 8N) increased

levels ofIL-IO transcript s. with a peak at 2 hours post treatment. The increase in RNA

express ion corre lated with an inc rease in IL- I 0 protein sec ret ion as measured by ELI SA.

Dose response curves showed that the level ofiL- IO protein stcad il y increased unt il it

reached a plateau at 222 nM of PSG 18N. To confirm that the increase in I L- I 0 secret ion

was not the result of priming macrophages with contaminat ing infl ammatory agents such

as LPS, induction of IL-IP, TNFa. iNOS. and IL-12 mRNA express ion was also

examined. PSG I 8 treatment did 11 0t increase express ion of these cytokines, indicat ing

that the induction of IL- I 0 was a specific effect. The results of these experiments suggest

that the biological effcct of PSG 18 is simi lar to human PSG 1. PSG6 and PSG II [ 12J and

more cruc ial func tions for PSGs may yet be found. Whether induction of macrophage

cytokine secretion is a function of other murine PSGs, or whether murine I>SGs havc

individual specia li zed fu nct ions has not yet been studied.

A Murine Model for PSG Function

Although CEA family members have 50-60% sequence similarity, identification

of rodent and human orthologucs for each member of the PSG subfamily has not been

possible. Although this complicates the eva luation of data from a murine model of PSG

function. investigation of PSGs in vivo and at the feta l leve l (for obvious ethica l reasons)

10

must be carried au( in a well-characteri zed animal system. Both human and mice are

hemochorial placentation mammals: maternal blood does not come into direct contact

with fetal blood duc to a layer of chorionic ti ssue surrounding fe tal blood vessels.

Therefore, murine PSG studies can indicate potential ce ll interact ions and reveal putative

ligands for thesc prote ins, which could greatl y facilitate understand ing or the function of

human PSGs. Furthennore, human PSGs have been shown to interact with murine

macrophage ce ll lines producing similar biologica l effects and suggesting functional

conservation between these species [12, 57]. Important ly, the complex steps involved in

embryonic implantation and development have been more thoroughly defined In mice

(reviewed by [64]), making these animals an ideal choice fo r study of PSG biological

function.

Pregnancy Specific Glycoproteins in Other Animal Species

In addition to humans and mice PSGs have been identified in the rat [65, 66], cow

[67], monkeys [32]. chimpanzee, orangutan [68]. and most recently, baboon [69] .

Streydio el al (1990) ana lyzed PSG gene development and concluded that the evolution

of the PSG gene family appears to coincide with the di varication of mammals, thereby,

coordinating placental development with expression of PSGs [20]. The similarities

between CEA proteins and genes in rodents and primates have been reviewed [27]. Of

particular note, all expressed PSG/CEA members across species contain an identical

number of amino acids in the leader peptide. Outside of this region, interspecies

comparison between PSGs has been more difficult. Computer analyses indicate that a

parallel but independent evolution of PSGs occurred as these two mammalian orders,

II

primates and rodentia, di verged perhaps resulting in the diffe ring numbers of IgV and

IgC-l ike domain, and RGD-like mOlifs observed in mouse and human PSGs today [20,

27] .

Placental PSGs: Expression of human PSGs in normal pregnancy and in cancer

During pregnancy, human PSGs are produced by syncytiotrophoblast [39], a

multinuclear syncytial cell layer or thc placenta. which separates fe tal and materna l

blood. III vitro, human embryo cultures secrete PSGs from three to four days post

fertilization [2]. Because of this early expression, PSGs are suspected to be invo lved in

trophoblasti c lIwasion of the uterus [14]. In vivo. PSGs are detectable in the maternal

serum seven days post conception. Blood samples obtained at two-week intervals over

the course of pregnancy revealed that PSG scrum concentration continues to rise

throughout the forty weeks [70]. It is interesting to note that human pregnancy specific

glycoproteins are the only pregnancy rclated proteins whose concentration continue to

ri se throughout pregnancy until birth [3]. PSG serum concentration reaches 200-400

[.tglml at teon l3]. The half-life of these proteins in maternal blood after parturition is

longer than other pregnancy related proteins at a value of 20-40 hours [71]. The high

serum concentrat ion and long half- li fe indicate that the serum concentrat ion of these

proteins returns to base line levels within weeks of childbirth [70] .

Evidence supporting the necessity of placental PSGs can be implied from the

observation that PSG maternal serum concentration, significantly lower than the median,

is predicti ve of spontaneous abortion f8]. Low maternal serum concentration of PSGs

correlates with intrauterine growth retardat ion. preeclampsia and fetal hypoxia [6, 7. 72].

12

Furthermore. a chronic reduction in expression of PSG 11 transcripts is observed in

endometria l ti ssue of women who suffer from recurrent spontaneous abortion [10]. The

importancc of PSGs during pregnancy is al so manifested in experiments using an imal

mode ls. Injection of anti-human PSG antibodies into monkeys induced abortion in these

an imals [73]. Spontaneous abortion also occurred when pregnant mice were injected

with rabbit anti -murine PSG antibodies l74].

Excess ive serum levels ofPSGs are associated wi th choriocarc inoma. invas ive

mole, hydati fonn mole (PSG6) and ovarian cancer [46, 75-77]. Serum PSG levels are

used clinica ll y to fo llow the progression of gestational trophoblastic diseases. PSG

quanti tat ion assists in detection of ectopic pregnancies. In these cases. PSGs are

identified in matelllal serum earli er than normal , before seven days. but at significantly

lower levels [78]. In normal pregnancy, maternal serum contains multiple spec ies of

PSGs, however it is not known whether different protein species are co-expressed at

similar levels throughout pregnancy, or if there is a pattern coordinating development

with express ion of different PSG proteins [38]. Two groups, Wu el at (1993 ) and

Chamberlain el al (1994) reported differentia l express ion of human PSG transcripts by

the placenta using RT-PCR and primers spec ific for the variolls PSG cDNAs [54, 79].

However, it is difficult, ifnot impossible, to correlate their mRNA expression findings to

protein expression, due to antibody cross-reactivity between PSG proteins.

13

A Brief Overview of the Multifaceted Phenomenon of Fetal Tolerance during Pregnancy

The fe tus expresses both maternal and paternal genes. and therefore exhibits nOI1-

materna l ce ll surface antigens. From this point of view. the comparison of the fetus to a

semi-allogeneic graft [80] is indispu table, howcver. controvcrsy stems from contentions

over the mcchan isms requ ired by the maternal immune system to mai ntain or tolerate the

"fetal allograft ," Medwar and Billingham proposed the following four hypotheses 10

explain the lack of fe tal rejection : the fetus is non-immunogenic. the conceptus represses

maternal immune responses, the uterus is an immune-privileged site, and the placenta

generates an anatomic immune panition between the mother and the unborn child [81].

Expcrimental evidence has already refu ted the first three hypotheses. Al though limited,

the feta l placental uni t does have immunogenic propenies [82-84J. The maternal immune

systcm is not repressed during pregnancy as women can mount an immune response to

infect ion by various pathogens. Introduction of foreign skin gra fts into the utcrus of

laboratory animals resulted in graft versus host symptoms and revea led that the uterus is

not an immune privileged si te as the grafts were infiltrated by maternal immune ce ll s

[85]. Several mechanisms thought to playa role in the regu lation of "fetal immune

privilege" arc rev iewed here. All discussed tolerance mechanisms can be ascribed to the

founh hypothesis. The first mechanism is described in more deta il. as it is the one that is

most directl y linked to current known PSG biologica l fu nctions. Indiv idual immuno-

regulatory mechanisms during pregnancy may function alone but it is likely that several

mechanisms co-exist. and should thus be viewed as a cluster of overl apping events

necessary fo r fetal surv ivaL

14

Anti-inflammatory Th2-Type Immune Response During Pregnancy as a Tolerance Mechanism

Both placental and uterine ti ssues produce and respond to cytoki nes, and these

interactions can dra maticall y affec t pregnancy ou tcome. Placenta l and uterine

mac rophages arc a source of transfo rnl ing growth factor (TGF) p. intcrl eukin (IL)-6 and

tumor necrosis factor (TNF) a [86]. Uterine ce ll s can generate TNFa [87]. TG Fp [88].

and granu locyte colony stimulating factor (G-CSF)[891. The uterus expresses IL-6 after

implantation [90] , 1 L-8 [9 1 J. and other cytokines during and aner implantation [92].

Trophoblasts produce both anti -inflammato ry and inflammatory cytokincs including IL-

IP [93][[94J. T Fa. IL-6 [95-97J, IL-8, IL-4, IL-I O [98. 99J and TGFp [94. 100l These

bioacti ve molecules can act locally in a paracrine or autocri nc manner. or at distal

systemic si tes regulating growth, development, and function of va rious ti ssues [ 10 I].

Cytokine production patterns by T helper (TH) ce ll s have been characterized into

three categori es: THO. TH I or T H2. THO cell s are precursor cells that can different iate

into either TI-I 1 or TH2 in response to signa ls from antigen presenti ng cell s (A PC). In the

presence of IL-4, THO ce ll s become TH2 ce ll s. which secrete IL-4, IL-5. IL-6. IL- I 0 and

TGF-p. Secretion of IL- 12 or IFN-y during antigen presentation results in production of

TH I cell s. These ce lls produce IL-2. TNFa. IL-12 and interferon (IFN)-y. IL- IO can

down-regulate the product ion of TH I cytokine synthesis while the adverse effect. an

inhibi tion ofTH2 cytokine synthesis. occurs in the presence of I FN-y [102]. The

products of both of these TH cell types can generate an immune response polarized

toward TI-I 1 or T H2, if large quantities of the cytokine are present. In the case of a

predominance o fTH2 cell products, an anti-in flammatory environ ment associated with a

humoral or ant ibody-mediated immune response is cultivated. In contrast, abundant TH I

15

cytokines stimulate a primarily inflammatory immune environment or ce ll -mediated

Immune response.

Precise patterns of cytokine express ion are critica l in regulating immune system

cel ls. A TH I immune environment has been shown to be detrimental to pregnancy in

both mice and humans. IL-2, TNFa and IFNy play important roles in abortion.

Injections of these in flammatory cytoki ncs terminate normal murine pregnancies [ 103-

106]. Spontaneous abortion can be induced with LPS which upregulates production of

inflammatory cytokines [107]. Additiona l evidence associating TH I immune responses

with hann to pregnancy is observed in studies of a mouse model of recurrent spontaneous

abortion (RSA). Abort ion occurs more frequentl y in CSA/JXDBA/2matings (50%) in

comparison to nonnal (CBAIJ X BALS/c) matings, where there is less than seven percent

fetal loss [ 108]. The reason for fetal resorption in thi s mating combination is unknown,

but is believed to represen t immune mediated reject ion of the semi-allogeneic fctu s,

because activated NK ce ll s infiltrate the placenta and destroy the developing embryo. In

addition, maternal immune cells from these matings respond to stimuli provided by

placental antigens i1l vitro with production of the inflammatory cytokines: IFNy, TNFu

and IL-2 [ 109]. RNA levels ofTNFu, IFNy and IL-2 were significant ly higher in

placentas from micc of the abortion prone mating combinat ion in compari son to nonnal

matings (11 0]. Moreover, injections ofanti-TNFu or anti -IFNy ant ibody increased

embryo surviva l [103] supporting the view that in fl ammatory cytoki nes are detrimental to

pregnancy. CBAIJ X DBN2 placentas produce less anti-inflammatory IL-4 and IL- J 0

than to controls [104). Embryonic rejection may also be directl y related to the increased

concentration of yo T cells present al the feta l maternal interface in abortion-prone

16

females [ I II]. Addition of anti-y chain an tibodies resulted in reduced abortion rates in

these animals [ 112]. More recently, in the murine model ofRSA. abOilions have been

blocked effectively by antibodies to fg l2 prothrombinase, suggesting that rejec tion is

mediated by a cytokinel thrombocyt ic mechanism [1 13]. Deciduas in abortion prone

pregnancies are characterized by an increased infiltrat ion of mate mal macrophages

[114]. In an inflammatory immune environment , these cells may destroy feta l ti ssues

through production of nitri c oxide (l i S]. It is uncertain which cell type is responsible for

the mitiatioll ora TH I immune response in the RSA mouse model but it is clear that

maternal immune cells from abortion prone females are capable of produce cytokines that

result in pregnancy tennination.

In humans. at least half of the cases of spontaneous abortion can be attributed to

factors that arc genetic. honnonal . anatomical or infectious [1 16] . The remaining cases

are proposed to result from maternal immune rejection of the fetus for unknown reasons

[117]. A diagnosis ofRSA syndrome or unexplained recurrent abortion (URA) refl ects

the clinical presentation of three or more successive miscarriages in the first trimester. in

which none of the previously mentioned causative factors could be identified. RSA

occurs in about 0.8 to 1% of pregnancies [ 11 8]. As in the mouse abortion-prone model.

evidence suggests that inflammatory cytokines may play an important role in human

spontaneous abort ions. During nonnal human pregnancy, a decrease in the type I

cytokines, IL-2 and IFN-y, in serum has been observed [119]. In contrast, studies

utilizing serum from women with URA found significantly higher concentrat ions of IFN

y, TNFu and TNFJ3 in comparison to controls [120]. Co-culturing of peripheral blood

mononuclear cell s (PBMC) from these women with auto logous placental ce ll s or antigens

17

from a trophoblast·deri ved cel l li ne resulted in higher levels of IFNy and rela tively low

levels of IL·6 and I L·I 0 production [12 1] . Immune system ce lls are highl y prevalent in

the pregnant uteru s. Vi nce et al demonstrated that T cell s and maerophages compose

approximately 20% of decidual ce ll s in the first trimester [89]. An increased number of

act ivated T ce lls localizes to the decidua in women with RSA [1 22]. Another group

spec ifica ll y showed that T ce ll s derived from RSA women exhibit a significant decrease

in IL·4 and IL·IO production [123]. A subset of women with RSAwas found to be

deficient in crit ical TGF·~·secreting suppressor ce ll s [ 124). Furthenl1ore, TNFa

treatment of human trophoblastic cells resul ts in apoptosis [125]. Collect ive ly, these

resu lt s indicate that TH I immune responses are just as are detrimenta l to pregnancy in

humans as they arc in mice [126]. Interestingly, Amold et al demonstrated that

monocytes derived from RSA women showed a significant decrease in I L· J 0 expression

level after treatment with PSG II [1 0] .

Experimenta l evidence li ke that described above has led to the development of the

theory that one important mechanism necessary to maintain the feta l allograft involves a

shift in the matemal immune response during pregnancy towards a TH2-likc environment

[127]. This theory was postulated based on the levels ofTl-l1 and TH2 cytokines in mice

during pregnancy and their effects on the matemal immune system and the outcome of

pregnancy. Anti. inflammatory cytokines are produced in part by the placenta , suggesting

that the fetus itse lf participates in creation of the appropriate immune environment for

successful pregnancy. For example, express ion of mRNA for the anti ·inflamrnatory

cytokine IL-4 by murine placenta is 5·1 0 fold higher than in peri pheral blood (128].

Murine placental tissues spontaneously secrete (LA, IL-5 and IL· IO, while maternal

18

lymph nodcs and spleen do not [129]. Immunohistochcmical staining of murine placental

ti ssues (day 10.5) revealed that a large quantity of the TH2 eytokines, ILA and IL-IO, are

produced by trophoblasts [130]. In general, I L-I 0 is known to downregulate the

production ofinflammalory cytokines. There is evidence that TI-12 immune responses

can thwart abort ive tcndencies. In the mouse model of RSA. reject ion can be prevented

by treatment with exogenous I L-I 0, suggesting that attack on the fetus is mcdiated by the

presence of TI-I I cytokines [1 04]. Each of these examples demonstrates the importance

ofTH2 cytokines during murine pregnanc y.

Less infonnation is avai lab le about human cytok ine production throughout

pregnancy. Phytohaemaggluttinin (PHA) stimulated PBM Cs from pregnant women

showed an increase in production of I L-4 and I L- I 0 [11 9]. As in the mouse. tenn human

placenta l tissues secrete IL-I 0 [13 1, 132]. Furthermore, placenta from earl y electi ve

abortions (weeks 8-10) and term placenta both sta ined pos itive by in silll hybridizat ion

for IL-IO. IL-4 and IL-3 mR NA, suggesting placental secretion ofthcsc cytokines [99.

130]. It is intriguing to note that IL-IO mR NA is intensely loca li zed in early invas ive

cytotrophoblasts yet it is on ly weakly present in these cel ls at tenn suggesting that IL-l 0

plays an essential role in pregnancy maintenance and poss ibly invasion but is repressed

prior to parturition [130]. Co-culturing of nonnal maternal PBMC with placental

an ti gens has demonstrated higher levels oflL-6 and IL- IO secretion than PBMC from

women with RSA [121 ]. Therefore, secret ion of anti -inflammatory eytokines is

associated with successful pregnancy in humans.

Clinica l correlation of the TH2 shift during pregnancy is derived from the effects

of pregnancy on materna l autoimmune di seases (rev iewed by [133, 134 J and infectious

19

diseases. During pregnancy, di seases associated with cell-mediated pathologies such as

rheumatoid arthriti s, autoimmune th yroiditi s and multiple sclerosis appear to follow a

pattem of remiss ion [1 35]. This effect is only temporary and symptoms retum atier the

birth orthe child . Conversely, Systemic Lupus Erythematosus (SLE), auto-immune

thrombocytopenia (ATP) and. autoimmune hemolytic anemia, aU humora ll y mediated

auto immune di seases. can be exacerbated during pregnancy [136]. Symptoms associated

with infections by intracellular pathogens such as Toxoplasmis gOl1dii and PlasmodiulI/

malariae are augmented during pregnancy. Normal resolution of these di seases occurs

through activaty of the cell-mediated immune system 1137]. Thus. the severity of

autoimmune symptoms varies with the pregnancy- induced immune environment, and

defects in the resolution of intracellular infections support the TH2-type shift paradigm.

Although both TH I and TH2 cytokines are expressed over the course of

pregnancy, the increased production ofTH2 cytokines may counteract the potentia ll y

harmful effects of inflammatory cytokines at the fetoplacental junction and systemically

through the innate immune response [1 26]. In addit ion, as stated by Wegmann

"alternatively, a TH2 response leading to antibody production may he useful to the fetus

since in many species these antibodies can be transferred across the placenta, resulting in

protective immunity for the neonate before the endogenous immune system develops"

[127]. The precise in vivo signaling molecules used to generate the TH2 bias are

presentl y unclear. This bias may be brought by hormones such as progesterone. relaxin,

corti sol and estrogen, and also by cytokines produced by macrophages, decidual cell s.

NK cell s, T lymphocytes, and trophoblast cell s [126, 127, 138- 140]. Other proteins such

20

as PSGs. the focus of our studies. and growth factors li ke GM-CSF and CSF. may also

contribute to the regulation of the immune response during pregnancy r I I, 12, 141 ].

Other Potential Mechanisms of Pregnancy Maintenance

In addi tion to a TH2-like immune environment. other potent ial mechanisms

involved in fetal prOiection include HLA-G express ion. honnone secretion, indoleaminc

2.3 dioxygcnase (100) synthesis. CD95/CD95L mediated apoptosis. and reduced

complement activation. HLA-G is a class ' MHC molecule expressed on the surface of

syncytiotrophoblasts in place of conventional MI-IC class I molecules [ 142] . In the

absence of class ical class I MI-l C molecules, such as HLA-A and I'ILA-B, at the fetal

maternal interface. syncytiotrophoblast may be able to escape killing by NK cell s via

express ion of HLA-G. HomlOnes. including progesterone and human placental growth

homlOne, have been reported to induce TH2 cytokine production and L1F expression

{143]. Indoleamine 2,3 dioxygenase (100) has been shown to be essent ial for the success

of murine pregnancy [144]. 100 is a tryptophan cataboli zing enzyme that is expressed

by macrophages and human syncyt iotrophoblasts [145,146]. and is postu lated to target

metabolica ll y active immune cells that are dependent on the availabi lity of tryptophan.

Pregnant mice treated with an 100 inhibitor di splayed activated T ce ll rejection orthe

fetus. 100 production has also been reported in human dendritic cells [ 147] and murine

endometrial ce ll s during implantation [148] , although it is unknown what triggers 100

product ion by these ce ll s. CD95/CD95L (Fas and Fas ligand) initiates a we ll -studied

apoptotic-signaling pathway. Tissue expression ofFas ligand (FasL) is associated with

immune privileged sites. such as the eye and testis . It has been proposed that

21

trophoblasls express FasL, which could kill activated immune cell s [149]. However

experiments with genetically Fas-deficient mice revealed that this pathway may not be

essen tial for fetal protection [ 130]. The imp0l1ance of regulation of complement

activat ion in protection of the murine fetus has been reviewed by Morgan and Holmes

[150] . The role of complement during pregnancy was discovered us ing Crry knockou t

[TIl ce . CtTy is a complement regulator in mice, and absence of thi s molecu le resu lted in

embryonic death due to extens ive C3 deposition in the placenta [15 1].

CD9 is a member of the Tetraspanin Superfamily

C0 9 is a member of the tetraspanin superfamil y of genes. Two independent

groups ident ified the first tetraspanin (CD81) in 1990 [152, 153]. To date there are 24

members of the tetraspanin superfamily [ 154-156]. A brieflisl of close ly relaled

Ictraspanin fami ly members includes: CD9, CD37, CD53, C063, CD8 1 and CD82.

Some members of thi s family are expressed in a wide range of ti ssues (CD9, CD63,

CD81, CD82), while others are cell -specific , like CD37, which is only expressed by

mature B cell s [ 154]. Tetraspanins are found in a variety of mammalian species and even

on the surface ofschistosomes (SM23)[ 152]. The diverse spec ies express ion of these

proteins suggests that they play key biological roles, which have yet to be identified.

Tetraspanin protein sequences predict structures with four hydrophobic transmembrane

domains, one small and one large extrace llu lar loop (EC I and EC2) and short

cytoplasmic domains at the C and N termini (figure I).

22

nlf Wl

CD9

< NH, COOH

CD37

-Jh mf»~~ fm lW~ <;j ~ lW

NH, COOH

CDSl

COOH

CD53

Ifff lW

COOH

Figure I. A schematic representation of the probable membrane orientation of four members of the transmembrane 4 superfamily (TM4SF). Potentially N- linked glycosylation sites are indicated by closed circles. (This figure was reproduced from Wright and Thompson. 1994. The Ins and Outs of the Tetraspanin Superfami ly. Immllflology Today 15 p.59 1.)

23

Amino acid sequence alignments of fa mily members can be viewed on the tetraspanin

homepage: ww\V.ksu.edu/tctraspan!thepagc.htm. (or [ 157]). There have not been any

li gands identified for the tetraspanin superfamily.

Because of their potentially diverse roles in biological processes, tetraspanins

have been described as "molecul ar facilitators"[ 154]. They mayfunction in regulating

cell deve lopment. acti vat ion, growth. and motility [158. 159]. Without a known

biological ligand for these proteins, much of the data regarding thei r fun ction has been

accumulated through use of anti-tetras pan in antibodies, comparison oftelraspanin

expression in normal and cancerous tissues , and the coimmunoprecipatation ofmolecu lcs

that associate with tetraspanins on the ce ll surface . The lauer has revealed a complex

network of protein interactions between a number oftctraspanin fam il y members and

integrin molecules including a3~1 ' 04PI , and a(>~1 [1 60- 164].

CD9 is a 24-27 kOa glycoprotein, which is widely expressed on hematopo ieti c

and non-hematopoietic ti ssues [165 , 166]. Sites of expression include platelets,

monocytes, basophils, eosinophil s, B cells, activated T ce lls. endotheli al ce ll s, neural

cells, vascu lar smooth muscle, cardiac muscle and epithelia [ 167- 174]. CD9 is also

expressed by non-T acute lymphoblasti c leukemia cel ls [ 175}, and in 50% of acute

myeloid and chron ic lymphoid leukemia [165]. In both megakaryocytic and pre-B cell

lines. C09 is associated wi th the tetraspanins C0 63, CD81 and C0 82. Work by

Rubinstein and Mannion showed that within thi s cluster, there is an association with the

MHC class II molecule HLA-DR and the very late antigen integri ns (VLA~I_ 4 and VLA

PI-6) [161 , 164). The authors suggest that the networking of these ce ll surface molecules

may indicate a role in signal transduction and ce ll motility. however the functional

24

significance of these interactions is still undefined. C09 has also been found in

assoc iat ion with the P2 integrin LFA-I (CD II /CD I 8) [176J.

A variety of antibodies to human and murine C09 have been used to elucidate

potential C0 9 functions. Treatment with ant ibodies to C0 9 have been shown to a lter the

di fferentiation ofhcmatopoietie ce lls. block production o f myelo id cell s in long-tenn

bone marrow cultures. and suppress proli feration of progenitor stem cells [177-179).

Some data obtained using ant i-C09 antibodies are contradictory. Ant i-C09 antibodies

have been reported to induce proliferation ofanti-CD3-act ivated murine T ce ll s. followed

by apoptos is due to insuffi cient IL-2 production by these ce ll s [ 180]. In contrast.

treatment of human CD9-transfectcd lurkat cells with anti-C09 and anti-C0 3 antibodies

resultcd in an increase in proliferation and production of IL-2 [181]. C0 9 expression in

various ce ll lines has been associated with heightened or repressed chemotaxis [182.

183]. Identi fi cat ion of the epitopes on C09 that are recognized by the different ant i-C09

antibodies, the C09 binding partner(s), and the intrace llu la r signa li ng pathways resulting

in these biological effects may explai n the diversity of reported functions.

Placental Implantation. Tetraspanins and Pregnancy

Immunohistochemical sta ining o1'C09 revealed that it is differentially expressed

on trophob last ce ll s [184] . Trophoblasts are specialized cells derived from the outer layer

of the blastocyst. Thcse cell s are responsible for implantation into the uterine wall. Once

partial implantation has begun, the trophoblast differentiates into two layers, an inner

layer of mononuclear ce ll s ca lled cytotrophoblasts, and an outer syncytial layer ca ll ed the

syncyt iotrophoblast. This layer is a true syncyt ium containing multip le nuclei and no

25

internal membranes to separate these organelles. In chorionic vi ll i, the

syncytiotrophoblast covers the cytotrophoblast cell layer (Figure 2). Cytotrophoblasts are

the "stem ce lls" of placental development. They can deve lop into honnone-secreting

villous syncytiotrophoblasts, ex tra-villous trophoblasts, or invas ive intermediate

trophoblasts. The honnones and proteins secreted by the syncytiotrophoblast layer.

which are critical for pregnancy maintenance, include human chorionic gonadotropin.

pregnancy specific glycoproteins, progesterone and estrogen. At the fetal-maternal

interface, where the placenta comes in contact with uterine ti ssue, extravi llous

trophoblast deve lops into anchoring cell columns attached to both the placen ta l villi and

the uterine decidua. Invasive trophoblasts are responsible for restructuring extracellular

matrix. These ce ll s seek out maternal arte ries, which they infiltrate. generating dilated

vessels that wi ll fi ll the intervillous space with blood from the mother. The maternal

blood wi thin the intervillous space provides for the exchange of nutrients, hormones and

gases between the fetu s and the mother. The pathological conditions pre-eclampsia,

eclampsia and intra-uterine growth retardat ion are all associated with abnormalities in the

trophoblast invasion process [64, 185].

26

placental seClum

ct\cnonic Pla le

I

leta l

veons

mlefvl!lous space , rnaln Stem

ma"~ l n ... c, rculallOn

ceclc ua p .... " •. '"

amr"cn

511,JmpOl m (l ln slem ",,1 us

C{lOlroonCbiaS IIC

,he'

Figure 2. Schematic diagram of the human placenta. Three chorionic vi lli extending from the main stem villus are disp layed in this schematic representation of the human placenta. The maternal portion of the placenta consists of the decidua basa li s. The fetal portion consists of the chorionic plate, chorionic villi and the cytotrophoblast! syncytiotrophoblast shell. Fetal vessels are lined by cytotrophoblasts which is directly covered by the syncytiotrophoblast layer. Maternal blood enter the intervillous spaces for the purpose of suppl yi ng the fetus with nutrients and oxygen and removing waste products. (Diagram reproduced from: Moore, KL. 1989. In Before We Are Born. Third Edition. Harvcourt Brace Jovanovich, lnc. PA, p. 92.)

27

Weak levels of C09 are expressed on the surface of ex tra vi llous trophoblast ce ll

columns and to a lesser degree on invasive trophoblasts (184). Syncytiotrophoblasts do

not stain positive for C09 expression at 7 weeks post implantation. In the second and

thi rd trimester, C09 is strongly expressed on ex tra-villous trophoblasts and chorion

laeve, which is derived from extravillous trophoblast, but not on either cyto trophoblasts

or syncytiotrophoblasts . in legrins n .l, 0::; and PI have also been iden tifi ed on extravillous

trophoblast [186]. Expression ofCD9 appea rs to down-regulate the metastatic potcntial

of these cells as it does in tmnsfected mclanoma ce ll s [182] and highly mctastatic cell

lines [182]. These findings \cd to the hypothes is that feta l CD9may have a regulatory

rol e in the invasion of cxtravillous trophoblast at the maternal-fetal interface.

Us ing a human CD9 eDNA probe, murine CD9 was cloned and its amino acid

sequence was predicted. 1·luman and murine C09 display 89% homology at the amino

aeid level [187]. To determine the physiological relevance ofC09, two groups

independently generated C09 knockout mice [188,189]. Although both male and female

C09-1- mice appear biologically normal, fema le C09-1- fema les exhibit a severe

reduction in fertility. In vitro fertilization experiments revealed that sperm wou ld not

fuse with oocytes from CD9-1- females. C09 is normally expressed on egg microvilli

and becomes clustered at the site of sperm attachment along with integrin u6PI and

fertilin [190]. Although the spenn attach nonnally, in vitro experiments demonstrated

that fusion of egg and spenn did not occur [188, 190]. interest ingly, when wild type

spenn were microinjected into C09-1- egg cytoplasm, the number of successful offspring

was equivalent to microinjection ofCD9+1+ eggs (189). Mating between male C09-1-

28

mice and wild type or heterozygous females resulted in nonnallevels of ferti li zation,

further indicat ing the necess ity of CD9 on the oocytes during fertilization [1881.

Specific Aims:

I. To study cytok ine expression of macrophages treated with the N·domain of PSG 17 .

2. To characterize PSG· macrophage binding

3. To clone the murine PSG Receptor

Significance:

The mechanisms. which generate maternal tolerance to the semi·allogeneic fetu s

are intertwined and essential for fetal survival. 1·luman PSG secretion begins within days

of conception and increases throughout pregnancy. Reduced concentrations of human

PSGs are presc ient of spontaneous abortions. Human PSGs have been shown to increase

cytokine production in macrophages. Each of these factors suggests that PSGs may play

a role in the network of fetal protection from maternal immunological response. This

study was designed to further investigate PSG immuno·rcgulatory mechanisms and

develop a mouse model of PSG function, including identification of the PSG receptor.

Hypothesis and Approach:

With the knowledge that murine PSG 18 and human PSGs: PSG I. PSG6 and

PSG II induce cytokines in maerophages. we will investigate the effects of the N-domain

of PSG 17 on a murine macrophage ce ll line and in thioglycollate·induced peritoneal

macrophages from BALB/c mice. Based on the highly homologous nucleotide and

amino acid sequences of thi s family and the similar biological effects observed with

29

human PSGs, we propose that murine PSG l7N will induce cytokine responses in these

cells.

Cloning orthe PSG receptor will allow us to identify o ther biological function s

for these proteins, and to ascertai n whether these prote ins share a common receptor.

Before choosing the cloning procedure to isolate the putative receptor, we sought to

biochemically characterize the PSG-receptor cell surface binding interactions and

identify the molecular weight of the interactive molecule(s). To detcrmlllc the affinity of

the PSG ligand for its putati ve receptor, binding curves for murine macrophages wi ll be

generated utili zing recombinant PSG-alkaline phosphatase fusion proteins. To

characterize the receptor(s) molecular weight, biOlinylatcd macrophage cell membranes

will be poured over PSG coated Nickel NTA beads (Qiagen). The protein(s) isolated by

the ligand· coated beads will be separated on an SOS-PAGE gel. transferred to

nitrocellulose and identified with strcptavidin-HRP. The protein(s) molecular we ight will

be determined based on standard molecular weight markers.

Results:

PSG 17N induced expression of IL-I 0, IL-6 and TGF-p I in both RAW 264.7 cel l

and peritoneal macrophages. T he concentrations of PSG l7N required to induce these

cytokines in RAW 264.7 cells were significantly less than that of PSG 18N. In add ition,

we determined that the PSG 17N increase of IL-I 0 and IL-6 mRNA in these cel ls required

de novo protein synthesis. Characterization of the PSG-macrophage ce ll surface binding

interactions revea led that there are approx imately 1770 PSG 17 bind ing sites per RA W

cell with a disassociation constant of2.2 X I 0-11 M. Based on the ability of PSG 17N to

30

induce cytokines in RA W cel ls and the high affin ity of the PSG 17 RA W cell binding

interaction, these cells were chosen as the RNA source to construct the eDNA expression

library for the purpose of cloning the receptor. Panning on PSG l 7N coated di shes was

utilized to sc reen the express ion library. In the work described here. it was dctennined

that CD9 is the receptor for murine PSG 17. In eluc idating the PSG 17-C0 9 interact ion.

we also di scovered the first biological ligand identified for any member of the tetraspanin

superfamil y. Therefore. CD9 appears to have more than one important func tion in

murine pregnancy: C09 is important fo r egg-sperm fu sion. it may regulate invasion into

the maternal ti ssues, and as a receptor for at least one PSG, it may infl uence the maternal

immune system.

31

II. Pape,· #1

Pregnancy Specific Glycoprotein 17 Binds with High Affinity to Murine Macrophages and Induces Production of Cytokines ill vitro.

Waterhouse, R., Wessells, .1., \"hire, K., Rollwagen, F., Zimmermann, \V., and Dveksler, G.

ABSTRACT

Pregnancy spec ific glycoproteins (PSGs) arc a famil y of sec reted proteins

produced by the placenta throughout pregnancy. In human serum, the concentration of

these glycoproteins thoroughl y exceeds that of other pregnancy related proteins, rcaching

lip to 200-400 Ilglml at term. Despite their high concen tration in materna l serum, no in

vivo biological function s for these proteins have been identified. In recent report s, we

have shown that treatment of cultured human clutriated monocytes with three

recombinant human PSGs induced IL-I 0, IL-6 and TGF-PI secretion . In addition, murine

PSGI8 was shown to increase IL-IO mR NA and protein expression in murine

macrophages, but its effects on IL-6 and TGF-PI secretion were nOl investigated. To

determine whether murine PSG family members have overlapping function s and mimic

human PSG biological effects, RAW 264.7 cells and thioglyco ll ate-induced peritoneal

macrophages were treated with murine PSG 17. We examined the leve ls of IL-l 0, IL-6

and TGF-PI secretion by ELISA after treatment with PSG l 7N a recombinant protein

cons isting of the N- domain of PSG 17. PSG l7N significantly induced secretion of al l

three cytokines. Treatment of RAW 264.7 ce lls with PSG 17N in the presence of

cycloheximide significantly reduced IL- I 0 and IL-6 mRNA levels indicat ing production

of these cytokines required de novo protein synthesis. Addit ional characterization of the

PSG-RA W 264.7 cell interaction was pursued with a recombinant PSG 17-alkaline

phosphatase fusion protein. Scatchard analysis revealed there were approx imately 1770

binding sites per cell with a K[) of2.2 X 10.11 M. The overlapping funclions o f various

PSGs and their high affinity to a ligand present on macrophages support the hypothesis

that these placentally secreted glycoprotcins arc involved in the regulation of the maternal

Immunc response.

I NTRODUCTION

Pregmmcy specific glycoprotcins (PSGs) arc a famil y of highly homologous

proteins secreted by the placenta. They were ori ginally iso lated frol11 the circulat ion of

pregnant women (I). PSGs are detccted in ma ternal blood as earl y as seven days post

implan tation. The serum levels of these prote ins reach up to 200·400 ~g/ml at term, far

exceeding the concentration of human chorionic gonadotropin and a lpha fe toprotein (2).

Abnonnally low levels of PSGs are associated with several serious compl ications of

pregnancy including fetal hypoxia, feta l growth retardation, pre-eclampsia and

spontaneous abortion (3·6). PSG homologucs have been identified in mice. rats.

monkeys, and most recently baboon (7·9) . Treatment with anti·PSG antibodies in two

animal models resulted in spontaneous abortion (10. 11). The high concentration of

PSGs in maternal serum and, their linkage to fetal pathologies and experimentally

induced abonions indicate that PSGs playa key role in pregnancy.

There are II human PSG genes (PSGI .PSGII) localized 10 chromosome 19.

Alternative splicing of these genes results in more than 37 different gene products.

Transcripts from these genes are co·expresscd during pregnancy. A peptide derived from

the N-terminal domain of human PSG 11 was shown to bind to macrophages. but not to B

33

or T cells (12). Studies with recombinant PSGs revea led that PSG I, PSG6. and PSG I I

induced the product ion of interleukin 10 (I L- I 0 ), I L-6 and transfo nlling growth factor /31

(TGFl3 d by human monocytes (13). Interestingly. activated monocytes from women

with recurrent spontaneo us abortion displayed a decrease in PSG I I-induced expression

of iL- IO ( 14). Together, these findings suggest that human PSGs are immune regu latory

facto rs. which can potent iall y affect maternal immune responses during pregnancy.

Partial genomic sequences have revea led that there are 14 murine PSG genes

(PSG/6-29) loca li zed to chromosome 7, which is sYlltenic to human chromosome 19 (8).

Murine PSGs arc similar to human PSGs in that they both contain several

1I11111l1noglobulin-like domains and are heav il y g lycosy lated. In both spec ies the N­

domain is immunoglobulin-variable like. The amino ac id sequences in this region are

highl y conserved between PSG genes within each species (54-94% in murine PSGs). To

date, full -length c DNAs have been reported for PSG 17, 18 and 19 (formerly Cea2. 3. 4)

(8, 15). Murine PSG 18 induced an increase in expression of IL-I 0 mRNA and protein in

a macrophage cell line and peritonea l macrophages (\5). Whether murine PSG 17 and

PSG 19 also induce IL-I 0 had not been reported.

In thi s study, we investi gated the induction of cytokine sec retion by murine

PSG 17 using a murine macrophage cell line and thiog lyco llate-induced peritoneal

macrophages. We also examined the binding kinet ics of the PSG-macrophage

interaction . Because murine PSG 17 induced IL-I 0, IL-6 and TGF/31 in murine

macrophages. we propose that murine PSGs have bio logica l functions s imilar to human

PSGs. In addition, characterizat ion of PSG-macrophage binding revea led that PSG 17

binds with high affinit y to a putative receptor on the surface of these cells. Together,

34

these results suppon a role for PSGs as immune regulatory molecules that could

poten tiall y influence matemal innate immune responses and feta l to lerancc.

MATERIALS AND METHODS

Allimols alld Cell Cil/illre

RA W 264.7 ce lls (an Abelson leukemia virus-transfonn cd macrophage cell li ne

of BALBlc origin) were obtained from American Type Cu lture Co llec ti on (Manassas.

VA) and were cultured in Dulbecco's mod ified Eagle's medium (DMEM) with hi l!h - -glucose, 5mM sodium pyruvate (Irvine Scientific. Santa Ana. CAl, 100 U/ml penici lli n.

100 ~glm l streptomycin, 0.25 ~glm l amphoteric in B (PSA) and 10% fe tal bovine serum

(FBS). HEK 293,. (PEAKRapid) ce ll s (Edge BioSystcllls, Ga ithersburg. MD) were

cultured in DMEM, 10% FBS. 50 ~g/ml gentamicin (Quality Biological Inc,

Ga ithersburg, MD), 250 ~glml G4 18 (CalBiochem, La Jolla. CAl and PSA. All ce ll

cultures were maintained at 37°C in a humidified atmosphere contain ing 5% CO2.

Five 10 six week old BAlB/c mice were purchased from the NC llaboratories

(Fredrick, MD). The animals were placed in cages with filter tops and fed standard chow

and water ad lib itum. Peritoneal macrophagcs were harvested and cultu rcd as previous ly

described (15).

Reagents

To gcnerate recombinant PSG 17N-Myc-His prote ins. thc N terminal domain of

PSG I7 was ampli fied by PCR from full length PSGI7 ( 16) ill pBluescript II KS+

(Stratagene. La Jo ll a. CAl using the oligonucleotide primers 5' GAAGATCTAGAG

35

ATATGGAG(T/G)TGTC3 ' (underlined Bg/11 si te) and 5' TTGGTACCCTCA TTT

(A/G)TCACAG(CIT) CAGG 3' (underli ned Kp" I s ite). The pe R product was inserted

into the pC RII .TOPO cloning vector ( Invitrogen. Carl sbad. CAl. To add the myc­

epilopc and 6X hi stidine tags to the N te nninus o f the e DNA. PSG 17N was removed

from pe RIl-TOPO by digestion wi lh Bgfll and Kpn i, and ligated into peONA3. 1 Myc­

His ( Invit rogen) . PSG 17N-Myc-His was inserted into the Not l and Psil di gested

pFastBac (Life Technolog ies. Rockville MD), Recombinam bacuJovims was obtained

followi ng manufacturer' s instructions (Life Techno logies. Rockvi ll e, MOl and purified

from insect ce ll supernatant as previously described for PSG 18N-Myc-His ( 15)1. The

GST -His-XyIE contro l protein was puri tied as described (1 5) using glutathione-sepharose

beads. All of the recombinant proteins were tested for endotoxin contamination using the

Limllills amebocyte lysate assay (Biowhittaker Inc .. Woburn, MA). On ly endotoxin free

recombinant prOieins were used in experiments measuring cytokine production .

Sllbcioning of PSG l7FL into Alkalille Phosphatase (A P)-Tag4

To generate AP-PSG 17, full length PSG 17 was amp li fied from PSG 17-

pBluescript 11 KS using the sense primer 5' GAAGATCTAGAGTCACTGTGGAATT

3'(underlincd Bg lJI s ite) and the anti-sense primer 5' TCACACGATCATCACAGCCAG

3', and cloned into pe R Script KS+ (Stratagene). The cDNA was excised with Bg lI! and

Xho l and inserted into complementary s ites within the AP-Tag4 vecto r (generous ly

supplied by Dr. John Flanagan, Harvard Uni vers ity), which resulted in a fu sion protein

composed of the secretio n signal followed by an alkaline phosphatase and the PSG 17

coding sequence at the C-tenninus. Pro toco ls for generat ion and quanti tat ion of heat

I Thesis Foomotc: Reier 10 Appendix 2 for a dCI,lik-d explanation or PSG 17NM yc-Hi~ rillrifical ion ~chemc_

36

stable AP-fus ion proteins have been reported by Flanagan and Cheng (1 7). HEK 293T

ce lls were transfected with the PSG 17-A PTag4 veclQr using Lipofectami ne I)LUS (Life

Technologies). Cell supematants containing the A I) proteins were harvested 5 days post­

transfection. filtered through a 22 micron filter. and 10 111M HEPES and 0.1 % sodium

azide were added. The supemalants were heated to 65uC for 10 minutes to inactivate

endogenous phosphatases. Serial dilutions of harvested supematallts were mixed at a I: I

ratio with 2X Ai> substrate (12 mM p-nitrophenyl phosphate (Sigma Chemi cals. St.

Louis. MO), 0.5 mM MgC\::, I M diethanolamine [pH 9.8], 0.5 I11g/ml bovine serum

albumin , and incubated for I hour at room temperature. The concelllratioll of AP protein

was measured by de-phosphorylation of p-n itrophenyl phosphate which was quantitated

by absorption at 405 nm using an ELISA plate reader. Whcn high concentrations of

fu sion protein were necessary. supematants wcre concentrated in a ccntriprep-I 0 and re­

assayed before treating cells (M illipore, Bedford, MA). Dilutions of AP-proteins for

binding assays were made in HBHA buffer (Hank's ba lanced sa lt so lution: 0.5 mglml

BSA, 20 mM HEPES [pH 7.0] , 0.1 % NaN,).

AP-FliSiol1 Proteil1 Binding Assays

RAW 264.7 cell s were grown to confluence in a 6-well plate and washed wi th

cold HBI-IA buffer. AP-PSG 17 or the AP control protein was added to the cells in

trip licate we ll s at a designated concentration. After 90 minutes at 4"'C, the proteins were

removed, and the cells were washed seven times with HBI-I A buffer. CeJllysates were

harvested w ith lysis buffer (1% Triton X- I 00, I OmM Tris - HCI [pH 8.0]), thoroughl y

vortexed. and then centrifuged to pellet nuclei and debri s. The cleared Iysates were

37

transferred to clean tubes and heated to 65°C fo r 10 minutes. The concentrat ion of AP

and AP· PSG 17 proteins in the cleared Iysates were quantitated by enzymat ic assay as

indicated above.

Analysis of Cyrokine illdllClioll by PSG·lrellwd mllcrophages (lml RA W 264. 7 cells

RA W 264.7 ce ll s and peritoncalmacrophages we re seeded in 24 well tissue

cu lture plates at I X 101> or 1.5 X 10(> ce ll s per well, respec tively. The cells wcre treated

in triplicate with PSG 17N-Myc. I-lis for 4 hours at 37°C in a vo lume of 300 !-II. After 4

hours. the vol ume was increased to I 1111 in each well by addition of cell culture media.

Supernatants were harvested at 2, 6. or 24 hours pOSHreatment as indicated. Secreted IL·

10, TGF·~l ' PGE1 and IL·6 were measured by ELI SA (R & 0 Systems. Minneapolis,

MN) (Pierce-Endogen, Wobum, MA). The limit of detection of the lL·I O, IL-6 and

TGFPI ELISA assay was 7 pglml and the PGE2 assay was 36.2 pglml. Prior to

measuring IL- I a and IL-6 secretion by BALBlc peritonea l macrophages. the supernatant

was concentrated 2.5 fold with a m icrocon -YM3 concentrato r (M illipore). Treatmen t

with LPS was used as a positive control for these experiments.

Semi-qllanlillllive RT-PCR

RA W ce ll s were grown in a 24 we ll plate were treated with 15 ~glm l

recombinant PSG protein or the contro l protein GST-I-lis-XyIE and cyc lohex imide

(S igma) was addcd at a concentration o f 5 ~lglml where indicated . Total RNA was

halvested two hours post-treatment using TRlzol (Life Techno logies) according to

manufacturer' s instruct io ns. RNA was reverse transcribed using random hexamers and

38

Ready-to-go you-prime-first strand beads (Amersham Phannacia. Piscataway, NJ). One

tenth of the reverse transcriptasc reaction was used for PCR amplifications of I L-I O. IL-

6, or GADPH. Various PCR cycles were analyzed to optimize the linear correlation

between RNA and PCR product. Amplified products were electrophoresed on a 1.5 %

TBE agarosc gel and blotted onto Nytran membranes. The membranes were UV cross­

linked, baked for 30 minutes at 80°C, and then hybridized with an internal 3~ p_ labeled

ol igonucleotide probe. After hybridization , the band intensity was quantified using the

Storm Phosphorimager and Image QuaNT program (Molecular Dynamics. Sunnyvale.

CAl. IL-IO and IL-6 cDNA intensity measurements were normalized to the GADP J-I

peR product value (I S).

Data Ana(vsis

Data was obtained from at least three independent experiments. Result s were

eva luated for statist ica l sign ificance using the unpaired student's I test. Data was

expressed as mean ± standard error (S.E.) and significance was defi ned at p <0.05.

RESULTS

PSG /7 induces secretion of IL- I 0 ill lIlurille macrophages ill a dose dependent manner

We previously reported that 20-25 Ilglml murine PSG ISN was required to induce

IL- IO in RAW 264.7 cells (\5). To dctennine ifmurine PSG 17 and PSG 18 have simi lar

bio logical effects, RAW 264.7 cells were treated with increasing concentrations of

truncated recombinant PSG 17 protein consisting of the N-domain (PSG 17N-Myc-His) or

a control protein GST-His-XyIE. and IL-I 0 secretion was measured six hour post-

39

treatment. Immunoreacti ve IL-IO showed a dose dependent increase in secretion after

treatment with PSG 17N (Figure I A). PSG! 7N induced significant levels of cytokine

secretion at a concentrat ion as low as 10 Ilglm l. A time course ana lysis revealed that

PSG 17N treatment of RA W 264.7 cell s resulted in an up-regu lation of IL- IO mR NA

express ion within two hours post treatment (data not shown). Based on these da ta, we

investigated the ab ili ty of PSG 17N to induce Il - I 0 in thioglycollate- induced primary

peritoneal macrophages. Treatment with PSG 17N at a concentration of 25 Ilglm l resulted

in significantly increased secretion of I l - I 0 in comparison to control protein and

untreated ce ll s (figure I B). The amount ofll- IO induced by PSGI7N in peritoneal

m3crophages was less than the amount induced by the positive control, J 00 ng/ml lPS

(data not shown).

PSGs ind/lce production of /L-6 alld TGFfJ i in murine macrophages

Previously, we showed that recombinant human PSG I, PSG6 and PSG II induced

secret ion of Il- IO, Il-6 and TGFp , by human elutriated monocytes (13). To determine if

murine PSG 17 treatment would mimic these effects in murinc macrophagcs, we

examined levels of IL-6 and TGFI3 I secret ion by RAW cells, six hours after treatment.

PSG 17N induced significant amounts of IL-6 at a concentration of 10 Ilg/ml. and the

response was dose dependent (Figure 2A). IL- IO and 1l-6 secretion was observed onl y

after treatment of RA W cells wi th murine PSG 18N2 at concentrat ions higher than 20

~lglml (data not shown). In addition , peritoneal macrophages treated with 25 )J.g/ml

2 To Ix: included only in Thc~ i s pub lication : I'S(l I RN ;oouct iol1 o f 11.·6 was perto nned by Jennifer Wessel ls as part o f her gradual ~ Ihe~is work ( 1999) al Ihe Unifonned Services Ull;vcr:< ity o f the Ileahh Sciellecs t ll lnlcd · · ·Th~ Cloning. Cha racleri7.ation and FunClJ<)nal Ana lysis of 1I. Iun rlc I'rcgn:mcy Sf'I.'Cilk G lycor1\lIClll ~." · II is prese nted in Ihi s manuscn pl as;' contri bU1ing dalJ poin1 1o !he whole and should be >I' noted.

40

A.

B

300 - - - - - - _. - _ .. - - - - -- ._.--- -- - -

250 ~

E 200 -'" 8 150

-

I-+- GST-HiS-XY1 E i I-O-- PSG17N I

•

•

•

::::; : ~ ~ 1 ~: 1+~Q=~-----.-------r-------.------.-------.---­o 5 10 15 20 25

Trclltmcnt(flg/ml)

300

250

200

~ ~ 150

100

50

o Unl re-a l~d GST- lI j §-X~'I E PSG 17N

Figure I. PSG 17N induces fl - I 0 secretion in murine macrophages and RA W 264 .7 cell s. (A) RA W 264.7 ce ll s were treated in tripli cate with increasing doses of PSG 17N­Myc-His or the control prote in GST-His-XyIE. Cell supernatants were harvested six hours post-treatment. and IL- IO secretion was measured by ELISA. All data shown arc representative of at least three independent experiments (*p<O.05). (8) BALB/c peritoneal macrophages were treated in triplicate with 25 J..lglml PSG 17N-Myc-His or GST-His-XylE, and fl - ! 0 production was measured from concentrated supernatants by ELISA.

4 1

A

1000

800 -E 600 '" Q. -"" 400 ..i - 200

0 0

B

-+- GST- ll i~·X yIE

~PSG I 7N

5

•

10 15 20

TrL'atml'nt ( ~lg/ml)

Unll"Ultd GST- H i~ - PSCl7N XylE

•

25

Figure 2. PSG17N induces IL-6 secretion in murine macrophage and RAW 264.7 cell s. (A) RAW cells were treated in trip licate with increasing doses of PSG 17N-Myc- His or GST-His-XyIE. Supernatants were harvested 6 hours post -treatment. and IL-6 secretion was assayed by ELISA. (8 ) Thioglycollate- induced pcrilOneal macrophage from BALB/c mice were treated in triplicate with PSG I7N or GST-XyIE-His at 25 j1g1ml. All data arc representative of three experiments (*p<O.05),

42

PSG 17N showed a s ignificant increase in IL-6 secretion (Figure 2B). PSG ISN also

induced IL-6 secreti on in BALB/c peritoncal macrophage at 25 ~tglml (data not shown).

TGFP I is associated wi th increased production of IL-I 0 and is bel ieved to be an

essential immuno-regulatory cytokinc during pregnancy (19) . In RAW cells, PSG 17N

induced TGFPI sec retion in a dose dependent ma nner (Figure 3A), and sign ificantly

increased TGF(31 production in BALB/c peri toneal macrophage at treatments of 10 tJ.glml

(data not shown) and 25 ~lg/m l (Figure 3 B). Unlike what was observed in RAW ce lt s.

PSG 17N induced s ignifi cant levels of TGFP I prote in as eart y as two hours post treatment

in peri toneaimacrophages. but not at 24 hours (data not shown).

PSG inductio/1 (~lIL-6 and IL -/O requires de IIOVO prorein ~:v/'lrhes is

Since delayed induction of IL- I 0 and IL-6 (twelve hours and six hours,

respecti vely) is observed after PSG treatment of human monocytcs , we proposed Ihal de

novo protein synthes is is required for PSG-induced IL- IO and IL-6 up-regulation (1 3).

To evaluate thi s hypothesis, RAW ce lls were treated with PSG 17N in the presence or

absence of the translat iona l inhibitor cycloheximide. In the presence of cycloheximide,

there was a s ignificant decrease in IL- IO and IL-6 mRNA compared to treatment in the

absence of cyclohex imide (Figu re 4). Co-treatments of LPS and cycloheximide resulted

in up-regUlation of IL-JO mRNA in comparison to treatment wit h LPS alone (data not

shown).

43

A

1400 - ----------- - ---- --- --- --•

1200 - -.-GST-His-XyIE •

~ 1000 ----6- PSG I 7 N

Q, 800 ~

• • ""- 600 ,

~~~ j .~!- -~ '" c..l ...

• :+;

0 5 10 15 20 25

Treatment (pg/ml)

B

2SO

200

50

o+---UnlrUled GST· nis-X~·I E PSG I7'"

Figure 3. PSG 17N induces TGFp. secret ion by murine macrophage and RA W 264.7 cell s. (A) RAW 264.7 ce ll s were treated with increasing doses of PSGI7N-Myc-l-li s or GST-His-XyIE. At 24 hours post-treatment TGFp. was assayed by ELISA. Data is representat ive of four experiments (*p<O.05). (8) BA LB/c thioglycollatc-induced peritoneal macrophagcs were treated with PSG 17N 25 Jlglml. TGFP I was measured by ELISA in concentrated supernatants from macrophages 2 hours post-treatment. Data is representat ive of three independent experiments.

44

PSG 17N

PSGl7N + CHX IL-6 0 IL-IO i ,

, " "' ' 00 '" '"

Fold I"crcas£' O\'l'r Control

Figure 4, Cyc loheximide inhibit s PSG 17N induction of IL-6 and IL- I 0 mRNA, RAW 264.7 cells were treated wi th 10 ~g/ml PSG I7N-Mye-l-lis or GST-His-XyIE in the presence or absence of 5 ~g/l11l cyclohexi mide (CI-IX). RNA was harvested at 2 hours for IL-IO and at 4 hours for IL-6 and leve ls of IL- I 0, IL-6 and GADPI-I were detennined by RT-PCR. The values reported were normalized 10 the GADPH va lues for each treatment and subsequent ly di vided by the va lues for the GST-I-Ii s-X yIE con trol treatment.

PSG 17 iI/duces the eicosalloid. Pros faglalldill E1 ill RA W 264. 7 cells

Prostaglandins are produced in low leve ls throughout pregnancy and at high

levels during labor (20) and PGE2 can regulate productio n of lL- l 0 and IL-6 in murine

macrophages (21). Therefore , we decided to test for the presence of signa ling thi s

molecule in RAW cell supernatants a fter treatmen t with I 0 ~lglml PSG 17N. Significant

levels of this e icosanoid were produced in response to PSGI 7N (Figure 5). Supernatant

analyzed at two, four and s ix hours post treatment showed that PGE2 secretio n peaked at

two hours. Expression was not s ign ificant at 24 hours post treatment (data not shown) .

In contrast. thioglycollate induced peritonea l macrophages did not demonstrate a

signifi cant increase in PGE2 secretion at two hours post treatment (data not shown).

45

l rn~

9lm

"'0 700

~ 600

, '00

~

" ' rn' • JOIJ

~on

100

0

GST. ll i~-X~ I£ PSGI7N

Figure S. PSG 17 I induced PGE:z in RA W 264.7 ce ll s. RAW cells were treated in triplicate with 10 1lg/1ll1 PSG 17N-M yc- His or the control protein GST -His-XyIE. Supernatants were harvested 2 hours post-trea tment and the concentrat ion of PGE~ was detcnnined by ELISA.

PSG 17 billds with high affinity to a fimired nlill/ber o.(billdillg sites Oil RA W 264. 7 cells

Induction of cytokine and prostagland in secretion in macrophagcs by PSGs.

implies the existence of a macrophage· PSG receptor. To characterize the binding of

PSG 17 to RA W ce lls, we cloned the full -length PSG 17 eDNA into a vector containing an

alkali ne phospluuase (AP) tag. RAW cell s were treated in triplicate with increasing

concentrations of AP-PSG 17 fusion protein or AP only. Saturation of PSG 17 binding

sites was observed at 86 11M (F igure 6A). Conversion of the binding data to a Scatchard

plot revealed that there are an average of 1770 binding sites per RA W ce ll for PSG 17

(Figure 68) w ith a KD of approx imately 2.2 X 10.11 M (Figure 68).

46

A.

-; = ~ ., ~ ~ ~

B.

16 .

14

12 .

10

8

6

4

2

0 0 50 100 150

nM PSG I7

200 250

0.6,------------------, ,-. 0.5 · • = ;::; 0 .4

t 0.3 ' ~

~ " 0.2

~ 0. 1

O +----_---_---~~-~ o 0.005 0 .01

Bound 0.015 0.02

300

Figure 6. Scatchard analys is of PSG 17 binding to RAW 264.7 cells. (A) RA W 264.7 ce ll s were treated in triplicate with increasing concentration of AP-PSG 17 or the AP contro l. PSG 17 binding values were plotted after subtracting leve ls of non-specific binding by AP from tota l binding. Similar binding curves were obtained in th ree independent experiments. (8) Scatchard ana lys is of binding revealed an average of 1770 binding sites per cell with Koof approximate ly 2.2 X 10" 'M.

47

DISCUSSION

Beginning at conception, the fetus is a chimera of both paternal and maternal

antigens. Although the maternal immune system may recognize the presence of the scmi ­

allogeneic fetus. in most cases, a detrimenta l immune response docs not ensue. For this

reason, pregnancy is a mysterious immunological phenomenon. Based on experimental

evidence, various mechanisms, that are not mutuall y exclusive. are potentia ll y

responsible for fetal tolerance (reviewed by Thcllin er al.) (22). The most relevant, with

respect to the data we presen t here, is the deve lopment ora bias toward a TI-I2 or anti­

inflammatory maternal immune environment of the matcmal immune response during

pregnancy (23). A TH2 immune envi ronment is characterized by the presence of specific

cytokll1es including IL-I O. TGF-J}I and IL-4. These cytokines reduce the expression of

pro-inflammatory cytokines, which are known to be detrimental to pregnancy (24) .

Macrophages arc an essenti al source of cytokines during innate immune

responses. These ce ll s are affected by various c irculating factors produced during

pregnancy (25). In this study, murine macrophages were treated wi th one such factor

PSG 17, that is nonnall y secreted by the placenta. Usi ng recombinant PSG proteins, we

examined macrophage secretion in response to PSG 17, as well as ce ll binding

parameters. We ascertained that murine PSG 17 b inds with high affi nity to a putative

receptor on the murine macrophage cell line RAW 264.7. We also showed that the ill

vitro biological effects of two murine PSGs are s imilar. Both murine PSG 17 and PSG 18

induce sec retion of IL-ID and IL-6, albeit at different concentrat ions. Regulation o f

cytokine secretion by macrophages may be a critical function of these proteins as both

murine and human PSGs induced IL-IO, Il-6 and TGFp l.

48

Through induct ion of IL·IO, PSGs could con tributc to the down-regulation of

inflammalOry TI-II cytokines locally at the fetal materna l junction and/or systemically as

these PSGs circu late in maternal blood. IL· IO was essentia l in reducing inflammatory

responses in an imal models (26, 27). Increased concentrations of I L· I 0 have been

reported at the fetal-maternal interface in nomlal pregnancies in mice and humans (28).

Decreased concent rations of I L· I 0 in serum have been assoc iated with preeclampsia and

spontaneous abortion ( 14 , 29)«30). A con'elation between decreased endometrial

production of human PSG I I transcripts and recurrent spontaneous abort ion in women

has also been reported (14).

I L·6 secretion was induced by macrophages using similar concen trations of

PSG17N. IL·6 is a muJti ·functional cytokinc with the ability to stimulate a variety of

different cell s. IL-6 treatment induces T and B cel l growth and differentiation, and

enhances the differentiation of hematopoietic precursors. In association with TNFo. and

IL·1. IL-6 augments acute phase protein synthesis in hepatocytes, but it has also been

demonstrated to act as an important anti-inflammalOry cytokine both loca lly and

systemica ll y (31). Uterine dec idual ce ll s and macrophages have been reponed to express

IL·6 during implantation and throughout pregnancy in mice (32-34). Production of thi s

cytokine inhib ited the IL-l and TNFa production by macrophages (35). Based on these

reports, PSG induction of IL·6 by macrophages may also play important ro les in

placental regulation of maternal immune responses and in placental development. For

example, lL-6 constitutively secreted by trophoblast ce lls has a positive feedback effect

on IL·6 receptor expression, and induces production of human chorionic gonadotrop in

(HCG) from first trimester human trophoblasts in culture (36).

49

In our experiments, TGFPl secretion was up-regulated within 2 hours of PSG 17

treatment in thioglycollate-induced macrophages, however, significant induction of th is

cytokine was not detected at 24 hours in RA W cell s. The different kinetics in TGFpl

induction in thioglycollate-eli cited macrophages and the murine macrophage ce ll line

may reflect the differing activation states of the ce ll s. TGFpl increases lL-6 production

in peripheral blood mononuclear ce!ls, and up-regulates IL-l 0 production in murine

macrophages. which may explain the signiticance of early production of this cytoki ne by

activated macrophages (37). The relevance of PSG-induced TGFpl secretion by

macrophages to the later I L-I 0 and IL-6 up-regulation remains to be dctennined. In

contrast to IL-6 and IL-I a induction, as little as I ~glml of PSG 17 was necessary to

enhance the levels ofTGFp l. Human PSG I, PSG6 and PSG]I also induced significant

levels of IL-6, IL-IO and TGFpl at 24 hours post-treatment in murine RA W cells (13).

Together, this data indicates that certain murine and human PSGs may have similar . biological functions.

Secretion of IL-6 and lL- 1 0 by RA W cells after treatment with PSG 17 required

de novo protein synthesis. The proteins required to be synthesized prior to IL- IO and IL-

6 mRNA up-regulation are presently unknown. In an attempt to identify potential

inducers of these cytokines, we analyzed expression of the eicosanoid, PGEz and the pro-

inflammatory cytokine TNFu. (data not shown). PSG 17N induced PGEz in RA W ce ll s

within 2 hours of treatment. Interestingly, PGE2 regulates lL-6 and I L-l 0 production by

activated macrophages (21) . Secretion of PGEz has also been described in uterine

macrophages during pregnancy. PGE2 has been reported to regulate T cell

differentiation, favoring TH2 T-ce ll development through induction of IL- I 0 (38). The

50

levels of PGE2 measured at 6 hours post-PSG 17N treatment were similar to the va lues

obtained after treatment with the contro l protein (data not shown). This down-regulation

of expression at 6 hours might be explained by the presence of IL-I 0 in the media. I L-\ 0

has been shown to provide negative feedback for PGE~ secretion in macrophage and

placental ce ll s (39, 40). Recently, experiments with spleen ce lls from IL-IO knockout

mice revealed that endogenous I L-I 0 is a regulator of LPS induced PGE~ synthesis ill

I'ivo (4 1). To confirm that IL-I O secre tion was not a resul t of earlier macrophage

production of pro-inflammatory cytokines, we tested supernatants from PSG 17N-treated

macrophages for the presence ofTNF-a at two hours posl-treatment. PSG 17 did not

increase production of thi s cytokine. This concurs with our previous data from

macrophage treatments with human PSG 1, 6 and II and murine PSG 18, none of which

increased secretion of inflammatory cytokines at the RNA or prote in level. The absence

ofTl-ll type cytokincs in response to PSGs suggests that these proteins are not non­

specifica ll y activating macrophages, but are inducing a specific response (un like LPS,

which is a potent inducer of both TH I and TI-I2 cytok ines in macrophages.) PSGs may

be directly down-regulating the production of TH I cytokine transcripts, however it is

morc likely that the absence ofTI-11 cytokine production wou ld be indirectly mediated

via the effects ofTI-I2 cytokines.

Current ly the receptor(s) for PSGs is unknown. Identification of the receptor for

PSG 17 may assist in elucidating the signaling mechanisms that lead to the induction of

TH2 cytokines. Based on the ability of PSGs to induce similar biologica l effects, human

and murine PSGs may utilize the same receplor(s). Our data show there are an average

of 1770 PSG \7 binding si tes per RA W cel l with an approximate KD of2.2 X 10- 1 \ M.

51

Although the number of receptors for PSG 17 is low, the receptor- ligand interaction is of

high affinity. Binding experiments with AP-PSG 18N indicated a lower bindmg affinity

for thi s protein, and abou t half the number of receptors per ce ll compared to PSG 17 thus

it is unlikely that they utilize the same receptors (data not shown). The difference in

protein concentration required for cytokine induct ion by PSG 17 and PSG 18 also suggests

different affinities for putative receptors. The 32% difference in amino acid sequence

with in the N-domains of PSG 17 and PSG 18 may result in recogn ition of unique bind ing

partners for each of these PSGs.

A peptide derived from the N tenllinus of human PSG II. con taining the RGD

integrin-b inding motif. bind to monocytes, but not to B or T ce ll s (12). It was therefore

proposed that the biological function s ofPSGs involved binding of thi s motifin a manner

similar to other RGD containing integrin bind ing partners (42). However, human PSG I ,

which lacks thi s motif(but has similar KGD sequence). induced IL- IO, IL-6 and TGF~I

in human monocytes (1 3). Murine PSG 17 and PSG 18 contain an RGD-like RGE motif

at the conserved site within their N-domains. Both of these proteins also induced

cytokine secretion in macro phages indicating that the conserved RGD motif is not

essential for biological function. However PSG binding to it s putative rcceptor(s) may

yet tum out to be dependent on the presence of thi s mot if or similarly charged motifs

such as RGE and KGD.

This is the first report to characterize the binding affi nity for a murine PSG

protein . Binding reached a plateau when RA W ce ll s were treated with 86 nM AP­

PSG 17. We attempted to compare the binding affinity of PSG 17 for RA W cells and

peritoneal macrophages. Unfortunately. the binding curves obtained using peritoneal

52

macrophages were highly variable. All mice used in these experimen ts were BALBlc

females between six and eight weeks old. Differences in immune responses during

various stages of estrous have been reponed for BALBlc mice (43). These mice reach

sexual maturity between 40 and 60 days and the estrous cycle of these mice occurs every

4-5 days for approx imately 10 hours. The vari at ion in estrous cycle or sexual maturi ty

may explain the irreproducibility of the binding data.

While the total concentration of human PSGs in matemal se rum at term reaches

200-400 J.lglml. the physiological concentral ion of individual human PSGs is unknown.

The homogenei ty of thi s protein famil y prevents immunological ident ification of

individual fam il y members, because ant ibodies developed to human PSGs cross-react

with multiple fami ly members. There is no information availab le on the concentrat ion of

circulat ing murine PSGs because current ly there are no immunological reagents availab le

to identify murine PSGs. We are in the process of developing antibodies to recombinant

murine PSG proteins. which will contribute to the study of these immuno-regulators.

In conclusion, our results indicate that murine PSG 17 induces simi lar biological

effects in macrophages to murine PSG 18 and three human PSG fami ly members. PSG

functionalit y in macrophages, chromosomal locat ion, structural similarities, and a

common site of expression indicate that these genes have been highly conserved during

evolution. Their high levels of express ion in human pregnancy and decreased product ion

in fetal pathologies suggest a crit ica l role for PSGs in pregnancy. PSG antibody-induced

abort ion in animal models also contributes to thi s hypothesis. Induction of TI-12

cytokincs in macrophages by placentall y produced PSGs may reduce inflammatory

53

immune responses and augment the generat ion of an immune enviro nment compatible

with successful pregnancy.

REFERENCES

I. Bohn, H .. Arch. Gynokol. 210.440-57, 197 1. 2. Li n. T. M., Halbert. S. P. and Spellacy, W. N.,./. Clin. Ill vesl. 54.576-82, 1974. 3. Tamsen. L. , Acta ObStcl Gyncc:ol Sc:aJ1(163. 3 11 -5 , 1984. 4. MacDonald. D. J., Scott, J. M. , Gemmell , R. S. and Mack. D. S .. Alii J Obstet

GYlleco! 147, 430-6. 1983. 5. Silver, R. M., I-Ieybome, K. D. and Lesli e. K. K., Placellta 14.583-9. 1993. 6. I-I enz. J. B. and Schultz-Larsen. P., lilt .J Gynaeco/ Obstet 21 . 111-7. 1983. 7. Chan. W. Y .. Tease. L. A.. Bates. J. M .. Jr .. Borjigin. J. and Shu pert , W. L.. HlIlII

Reprod 3, 687-92. 1988. 8. Ruder!. F .. Saunders. A. M., Rebstock. S .. Thompson. J . A. and Zimmemlann.

W. , Mamm Genome 3. 262-73. 1992. 9. Zhou, G. Q. and J-Iammarstrom, S .. Bioi Reprod 64.90-99.200 I. 10. Bolm, I-I. and Weinmann, E. , Arch Gynokol 221 , 305- 12. 1976. II . Hau, J., G idley-Baird, A. A., Westergaard . J. G. and Teisner. 8. , Biomed.

Bioc:him. Acta 44. 1255-9, 1985. 12. Rutherfurd , K. J .. Chou, J. Y. and Mansfield, B. C .. Mol. Endocrinol. 9, 1297-

305. 1995. 13 . Snyder, S., Wessner, D., Wessells. J.. Waterhouse, R., Wahl, L. M.,

Zimmermann, W. and Dvekslcr. G .. AJRI 45, 205-2 16, 200 1. 14. Arno ld, L. L., Doheny, T. M. , FloI', A. W., Simon. J. A .. Chou. J. Y .. Chan. W. Y.

and Mansfield. B. C.,Am. 1. Reprod. Immullol. 41 , 174-82, 1999. 15. Wesse ll s, J.. Wessner, D., Parsells, R., White, K., Finkenze ller, 0 .. Zimmermann,

W. and Dveksler, G .. £lIr. 1. Immlll /Ol. 30, 1830- 1840,2000. 16. Thompson. J . A., Mauch, E. M., Chen, F. S., Hinoda. Y .. Schrewe. I-I. , Berling,

B., Bamert, S .. von Kleist, S., Shi vely, J . E. and Zimmemlann, W., Biochem Biopl/ys Res CommulI 158, 996- 1004,1989.

17. Flanagan. J. G. and Cheng, H. J.. Methods Ellzymo! 327. 198-2 10.2000. 18. Dveks1er, G. S .. Basile, A. A. and Dieffenbach. C. W. , PCR Methods Appl. I.

283-5, 1992. 19. Godkin. J . D. and Dore. J. J., Rev Reprod 3, 1-6, 1998. 20. Brown. N. L.. Alvi, S. A. , Elder, M. G. , Bennett, P. R. and Sullivan, M. 1-1 .,

Immunology 99. 124-33, 2000. 21 . Wi lliams, J . A. , Pontzer_ C. 1-1 . and Shacter. E .. J Illtel/eroll Cytokille Res 20. 29 1-

8.2000. 22. The llin, 0., Coumans, 8., Zorzi , W., Igout, A. and I-leinen, E., Cur,- Opin

IlIImllllo/ 12, 73 1-7, 2000. 23. Wegmann, T. G., Lin, H. , Gui lbert. L. and Mosmann, T. R .. lmmullol. Today 14,

353-356. 1993.

54

24. Raghupath y, R., Immullo! Today 18, 478-82, 1997. 25. Sacks. G. , Sargent. I. and Redman. C. , lm11/ul/ologv Today 20. 114- 11 8. 1999. 26. Berg. D. J., Leach, M. W .. Kuhn . R., Rajewsky, K .. Mu ller. W .. Davidson. N. J.

and Rennick, D., J Exp Med 182. 99- 108, 1995. 27. Chaouat, G., Meliani , A. A., Martal , J. , Raghupathy, R .. Elliot, J ., Mosmann, 1'.

and Wegmann. 1'. G .. .!. 111I1IIU1/0!. 154.426 1-4268, 1995 . 28. C haouat, G .. Cayol, V. , Mairovitz . V. and Dubanchet , S .. Am) Reprod Immlll/ol

42. 1-13. 1999. 29. Hennessy. A. , Pilmorc. 1-1. L. . S immons, L. A. and Painter. D. M .. ) I11111l11110! 163,

349 1-5, 1999. 30. Raghu pathy. R .. Makhseed. M .. Az izieh, F., Hassan. N .. AI-Azemi . M. and AI­

Shamali . E .. Cell 1111111 111101 196. 122-30. 1999. 3 1. Xi ng. Z., Gauld ic. J., Cox. G .. Baumann. I-I. , Jordana. M .. Lei. X.-F. and Achong.

M. K., .J. Ciill. IlIveSI. 101 , 3 1 1-320. 1998. 32. De. 111 .. Sanfo rd . T.R .. Wood, G. W. , J. Reprod l'el'li/ 97, 83-89. 1993. 33. Robert son. S. A.. Mayrhofer. G. and Seamark , R. F .. BioI Reprod 46, 1069-79.

1992. 34. Liang, L. , Kover. K .. Dey. S. K. and Andrcws, G. K .. ) Reprod 1111I11UI101 30, 29-

52, 1996. 35. Aderka. D., Le. J. M. and Vilcck, J. , ) IlIIl1Ilfl/o/ 143, 35 17-23. 1989. 36. Nishino, E., Matsuzaki , N. , Masu hi ro, K., Kameda. T. , Taniguchi , T. . Takagi. T ..

Saji , F. and Tanizawa, 0. , ) C/ill Endocrino! M elab 71 . 436-41 . 1990. 37. Turner. M. , Chantry. D. and Feldman n. M .. CylOkille 2. 2 11 -6 .• 1990. 38. Demeure , C. E .. Yang, L. P .. Desj ardins, c., Raynauld, P. and Delespcssf!, G. , E liI'

)1"'IIIltr1oI 27, 3526-3 1 .. 1997. 39. Strassmann, G., Pat il-Koota, V .. Finkelman, F., Fong, M. and Kambayashi , T.,)

Exp Med 180, 2365-70, 1994. 40. Goodwin, V. 1.. Sato, T . A. , Mi tchell , M. D. and Keelan, J. A .. Am J Reprod

IIIImunol 40, 3 19-25. 1998. 41. Berg, D. J ., Zhang, J. , Lauricell a, D. M. and Moore. S. A .. ) Il11l11u110/ 166, 2674-

2680, 200 1. 42. Ruoslahti , E. and Pierschbacher, M. D .. Science 238. 49 1-7, 1987. 43 . Krzych, U .. Strausser, H. R., Bressler, J. P. and Goldstein, A. L. , .J Imlll ullol lll.

1603-5. , 1978.

55

III. Paper #2

Murine CD9 is the Cellular Receptor for Pregnancy Specific Glycoprotein 17

Roseann Waterhouse and Gabriela Dvcks1cr Uniformed Services Uni vers ity of the Hea lth Sciences. Department of Patho logy

Abstract

Pregnancy spec ific glycoproteins (PSGs) are a fa mi ly of highly simi lar secreted

proteins produced by the p lacenta. During gestat ion, they become the most prevalent

pregnancy-related protein In maternal serum. Low concentrat ions of PSGs arc associated

with spontaneous abortion. It has been suggested that blood bome PSGs may act as

immune mediators protecting the fetus from maternal immune responses. I" vilro

experiments performed in o ur Jab, revea led that treatment of monocytes with human

PSGs increased sec ret ion of IL-I D. IL-6, and TGFPI. This data supports such a

hypothesis. providing an effective mechan ism through which immune protection could

occur. PSG homologues have been identifi ed in several species including primates and

rodents. Recently, we demonstrated that one murine PSG famil y member. PSG I 7, bound

to RAW 264.7 ce li s, a murine macrophage ce ll line, with high affin ity and induced

secretio n of anti -inflammatory eytokines in these ce ll s. Based on these findin gs, we

proposed that these ce ll s express the PSG 17 receptor. For the purpose of cloning the

recepto r, we sc reened a RAW cell cDNA express ion library by panning. The PSG 17

recepto r was identifi ed as the tetraspanin, C D9. We confinncd binding of PSG 17 to C D9

by ELISA, flow cytometry, alkali ne phosphatase binding assays, and in situ rosetting.

Treatment with anti-CD9 antibodies inhibited binding o f PSG 17 to CD9 transfected cells .

Two other murine PSGs, PSG 18 and PSG 19 were also tested for binding 10 C D9. They

did not bind to cell s transfeclcd with thi s tctraspanin, suggesting that PSG 17- C D9

56

bind ing is a spec ific interaction. Wc have iden tified the fi rst reccpto r fo r a murine PSG

as well as the first natmalligand for a member of the tetraspan in superfa mily. Wi th thi s

discovery, we can identify the s igna ling mechan isms used by PSG! 7 to induce cytokine

sec retion in mac ro phages.

Introduction

Pregnancy speci fic g lycoprote ins (PSGs) are a fam ily o f highl y s imi lar,

placentall y secreted protcins, o rigina ll y isolated from the circu lation of pregnant women

( I ). PSGs are members o f the Carcinoembryonic Antigen (CEA) famit y, which be longs

to the Immunoglobulin superfamily. In humans, PSGs are secreted into the bloodstream

and their concentnllio n increases exponenti ally until tcnn (2). Low levels of PSGs are

assoc iated wit h certain pathologica l conditions including spontaneous abortion,

intrauterine growth retardation , and pre·eclampsia (3). Injections of anti -PSG antibodies

induced spontaneous abortion in primates. indicating that PSGs are essentia l fo r

successful pregnancy (4) . Recent ly, we demonstrated that human PSGs induced IL·\ 0,

IL-6 and TGFf31 expression in human c1utriatcd monocytes (5). It has been suggested that

PSGs, through induction of these TH2 cytokines, may have a ro le in protecting the fetus

from attack by the maternal immune system.

Desp ite their isolat ion from maternal serum about thiny years ago, the receptor(s)

fo r these proteins has not yet been identified. Studies by Rutherford and co ll eagues

reported a pept ide deri ved from the N·lenllina l doma in of human PSG I I binds to human

mOllocytes and ce ll s of the promonocyte Iineagc. but not to T or B cells (6). PSGs have

also been identified in non·primates with hemochorial placentation includ ing ra ts and

mice. Screening ofa YAC library revea led that there are 14 murine PSG genes (PSG14·

57

29) (7). These genes were mapped to chromosome 7 (8) and the cDNAs thai have been

characterized showed exclusive express ion by the placenta (9).

We recent ly reported that murine PSG 17 and PSG 18 mimic the biological effec ts

of human PSGs. by inducing TH2 cytokincs in murine macrophages and RAW264.7,

murine macrophage cell li ne (10. II) . We also showed that PSG 17 binds to RA W ceJls

with high affi nity ( II ). Based on these findings, we proposed that these cel ls exprcss a

receptor for PSGs. For the purpose of c loning the putative PSG 17 receptor. we screened

a RA W cell cDNA express ion library by panning. Posi ti ve clones were sequenced. and

database queries revealed thaI the cloned binding partner fo r PSG 17 is the tetraspanin

C09. We con finned the specific binding of PSG 17 to CD9 using four d ifTerent

techniques and in Iwo different cc illincs. PSG 18 and PSG 19 were al so tested for binding

to CD9. They did not bind to CD9 suggest ing that the binding is a specific interaction

between PSG 17 and CD9.

CD9 is a member of the tetraspanin superfami ly. This molecule is expressed on a

wide variety of hematopoietic and non-hematopoietic cells. C09 associates with other

tetraspanin family members and integrins ( 12). To date. there arc no known ligands for

any of tile more than thirty members of the murine and human tetraspanin superfamily

(13. 14) (sec also: www.ksu.edu/tctntspan/thepagc.htm). Treatments wi th antibodies to

different tetraspan ins have implicated these ce ll surface proteins in cell migration.

proliferation, activation and adhesion {reviewed by (13». With such potentially diverse

roles. they have been described as "molecular facilitators." This is the first report to

identify the binding partner for a PSG fam il y member, as we ll as a biological ligand for a

member of the tctraspanin superfamil y.

58

Materials and Methods

Reagellt.,·

Recombinant PSG 17N-Myc-l-lis. PSG 18N-Myc-His, PSG 19N-Myc-His and

PSG6N-Myc-H is were generated using a baculovirus expression sys tem and purified

from insect cell supematant as previously described (5, 10, I I ). Ce ll transfcctions were

perfonned with Lipofectamine 2000 (Life Techno logies, Rockvi ll e. MO). PSG 17N­

Myc-l-Iis coated plates were prepared as described ( 15). Briefly, under sterile conditions.

bacteriologica l culture dishes (Falcon) were layered with 10 pg/ml goat anti- mouse IgG,

ex tra serum absorbed (XSA) (KPL Gaithersburg, MO) in phosphate coating so lution

(KPL). The plates were rinsed with PBS and blocked with bovine serum albumin (BSA)

buffer (KPL). An anti-myc monoclonal antibody [I pg/ml] (Invitrogen, Carl sbad, CAl

was added to each dish followed by washes. The recombinant murine PSG 17N-Myc-H is

proteins were subsequently added to the dishes.

Cell CII/ture

RA W 264.7 (American Type Culture Collection, Manassas, VA) cells were

cultured in Oulbecco's modified Eagle's medium (DMEM) with high glucose, 5mM

sodium pyru vate (Irvine Scientific, Santa Ana, CAl, 100 U/ml penic ill in, 100 pg/ml

streptom yc in , 0.25 ~lglm l amphoteric in B (PSA) (Quality Biological, Gaithersburg, MO),

and 10% Feta l Bovine Serum (FBS). HEK 293T (PEAKRapid) ce ll s (Edge BioSystems,

Gaithersburg, MO) were cultured in OMEM, 10% FBS, 50 pg/ml gentamicin , 250 pglml

G418 (CalBioehem, La Jolla, CAl, and PSA. EBNA 293 ee lls ( Invi trogen ) were grown

in OM EM, 10% FBS, PSA and 250 pg/ml G4 18. BHK- 21 ce ll s (generously donated by

59

Or. Gerardo Kaplan) were sustained in DMEM wi th 10% FBS and PSA. All ce ll cultures

were maintained at 3711C in a humidified atmosphere containing 5% CO2•

Generfllioll of a murine macrop/wge eDNA Iihrw)'

Total RNA was ex tracted from RAW 264.7 cells using TRlzol (Life

Technologies) as per manufacturer's instnlctions. Purification ofpolyadcnylated RNA

and generat ion of the RA W 264.7 ce ll cON A express ion library in the PEA K I OC V

vector was done by Edge BioSystems (Gaithersburg, MO). The unamplified li brary

yielded approximate ly 4.3 X 10(' primary transformants wi th an average insert size of

1700 bp and an insert cut-off of 700 bp. For the first round o f sc reening, 1.1 X 1 Of> clones

were plated on Luria-Bertani (LB) broth/agar plates with 100 j..tglm l ampicillin (Sigma,

St. Louis,MO). Aft er 16 hours the plates were flooded wi th LB broth, the pooled

bacteria were pc lleted, and plasmid DNA was isolated by ces ium chloride purification.

Recovery of eDNA clones by panning

Pooled purified plasm ids were transfecred illlo EBNA 293 cell s (Invitrogen) .

Positive transfectants were selected using 0.5 j..tglml puromyc in (Edge BioSystems). At

72 hours post se lection, the ce ll s were dislodged in PBS and 0. 5 mM EDTA (16), and

resuspended in binding buffer (PBS, 2% BSA). The detached ce ll s were panned in

PSG 17N-Myc-H is coated petri di shes at 1.0 - 1.5 X 107 cel ls per di sh (16, 17). Non­

adherent ce ll s were removed by washing with at least six PBS/BSA washes until there

were no floa ting cells visible under the microscope. Adherent cell s we re removed from

the bottom of the di shes by pipetting and seeded into in poly-L-Iysine coated wells.

60

Puromycin was increased to 1 J.lglml two weeks after the initial transfection. An

addit ional two rounds of panning were perfonned before the episomal plasmids were

iso lated using the H IRT extraction procedure. The plasmids were transfected into high

effic iency ElectroMAXTr-.·( 01-1 lOB ce ll s (Life Technologies) by electroporation. Twelve

random single co lonies were grown and the plasm ids purifi ed using Qiagen Midi-Prep

columns. Individual plasmids were transfectcd into 293T cells and the transfccted cells

were screened for binding of PSG 17N-Myc-l-lis by ELISA (see below). Insel1s from the

plasm ids that conferred PSG 17N binding were sequenced with AB I PRISM Big Dye

Temlinator Cycle Sequencing Read y Reaction Kits (PE Biosystems, Foster City, CAl.

Detection ofPSGJ 7 Binding to tralls/elud H£K 293T cells hy ELISA

For the identification of the PSGl 7 putative receptor, HEK 293T ce ll s were

seeded in poly - L lysine coated 96 well plates at 5 X 104 ce ll s per we ll and transiently

transfected with plasmid DNA obtained as described above. AI 48 hours post­

transfection , the cells were washed with binding buffer containing 0.0 I % sodium azide

and PSGI7N-Myc-His ( 10 J.lglml) (or no ligand) was added to each well. After I hour of

incubation at room temperature, the ligand was aspi rated and the cells were washed 5

times with binding buffer without sodium azide. To detect binding of PSG 17N-Myc-His,

anti-myc-horseradish peroxidase (H RP) conjugated antibody (Ab) was added to the ce ll s

for I hour at room temperature at a concentration of I ~lglml (in binding buffer). Binding

of the antibody to the ligand was detected after the addition of tetra methyl benzidine

(TMB)-peroxidase substrate (KPL) followed by 2N 1-I 2S04 (stop solution). The color

change was quantitated at 405 nm on a microplate spectrophotometer.

61

Competition experiments were carried out in CD9-pEF6/V5-l-lis (Invitrogen)

transfccted 293T ce ll s by adding increasing concentrat ions of anti-murine CD9 Ab (SO

Phaml igen, San Diego, CAl or an isotype matched con trol (rat IgG.:!a) for I hour at room

temperature prior to treatment w ilh 5 }lglml PSG 171 . Absorbance was normalized to

background binding of PSG 17N in the presence ofanti -CD9 an tibody to empty plasmid

transfecled 293T cells. To examine whether other murine PSGs bound to C09. ELiSAs

were also perfonned us ing PSG 18N-Myc-H is and PSG 19N-Myc- His at I 0 ~lg/m l in place

of PSG 17N-Myc- His.

/" Silll Rosettillg A .... \'ay

I II si/ll roselling assays modified from Kap lan el al. were perfonned to confinn

binding of PSG 17 to CD9 (18). Briefly, 293T cells transiently transfected wi th empty

vector or CD9-pEF6N5-His were seeded at low density into poly L- Iysi ne coated 60 mm

dishes (Becton Dickinson. Sparks, MO). 'Attached cells were washed in binding bu tTer

(PBS- 2% BSA). PSG 17N-Mye His (90 pieomolcs) or binding buffer alone was added to

the di shes fo r I hour at room temperature. The buffer was aspirated from each di sh and

the cell s were washed four times with PBS 10 remove any unbound ligand prior to the

addition of I }lg/ml an ti-myc Ab in bind ing buffer for I hour at room temperature. As a

control, the an ti-myc Ab was omitted from some plates. Dishes were washed again with

PBS prior 10 the addition of 15 }lg of rabb it anli-mouse immunoglobu li n coated beads

(BioRad, Richmond, CAl. Unbound beads were removed by washing extensive ly wi th

PBS and receptor positive ce lls were viewed by microscopy.

62

Floit' cytomelly:

I-I EK 293T or BHK-2 1 cells were transfected with murine CD9-pEF/V5-H is or

empty vector. Transfected ce ll s were di vided into polystyrene tubes at 5 X 10 5 ce ll s/tube.

Ce lls were washed twice with wash buffer (PBS: 3% FBS ; 0.01 % sodium azide), prior to

the addit ion of 10).lg PSGI7N-M yc-His for 30 minutes at room temperature. After two

washes, ce ll s were incubated on ice sequentia ll y with the following solutions for 30

minutes eac h: (see Table 1 & 2) 0.5 pg anti-myc Ab (Inv itrogen), 0.5 p g goat anti-mouse

I gG ~'1 kappa chain-biotin labe led Ab (BD Phannigen) and 0.5 ).l g streptavidin-conjuga ted

FITC (BD Phannigen) with two washes between each incubation. PSG 17N binding to

HEK 293T and BHK 21 cell s was analyzed by flow cytometry using an EPICS Xl-MC l

fl ow cytometer (Beckman Coulter, Miami, Fl) and the percent binding was detennined

with the System II Software program (Beckman Coulter). Overlays were produced with

the WinList program (Verity Software House, Topsham, ME).

Alkaline phosphatase binding assays:

PSG alkaline phosphastase (AP) fusion proteins were generated by cloning into

the AP-Tag4 vector generolls ly supplied by Or. John Flanagan (Harvard University).

C09 binding assays with heat stable AP-PSG fusion proteins were performed as

described by Flanagan and Cheng (19). 293T cell s transiently transfected with murine

C09 or empty vector were cultured in poly-L-Iys ine coated six well plates. increas ing

concentrat ions (43, 86 and 129 nM) of AP-PSG17 or the AI"' control protein were added

to each we ll in triplicate for 90 minutes at room temperature. The cells were thoroughl y

washed, and the concentration of bound protein was measured from cleared, heat-

63

inactivated ceillysatcs by a colorimetric reaction as previously described (I I). Binding

of AP-PSG 18N 10 CD9 trans fected cells was also invest igated al two concentrations. 8.6

nM and 86 11M.

Results

Identificati{)11 {)f CD9 as the cellular receptor j{)r PSG 17

Based on the ability of PSG I 7 to induce cytok ines in the mouse macrophage ccl l

[inc. RA W264.7, and the high binding affinity of PSG 17 for a ligand on the surface of

these cells, we proposed that these cells expressed the receptor for PSG 17 . To iso late this

receptor, a RAW ce ll eDNA library was generated (Edge BioSystems) in the PEAK I OCV

vector. Plasmid DNA from - 1.1 X 106 unamplified clones was Iransfecled into 293

EBNA cells and transfected ce ll s were se lected by addition o f puromycin, prior to

panning on PSG 17N-coated di shes. After three rounds of panning. plasmid DNA was

isolated by HIRT extraction and electroporated into ElcctroMAX DH lOB E. coli.

Plasmid DNA iso lated from individual bacterial co lonies was re-transfected into HEK

293T cells and their ability to confer PSG 17 binding was detemlined by ELISA. All

positive clones were sequenced. Database queries identified a perfect match between

these cDNAs and the tetras pan in family member, murine CD9.

To confirm binding of PSG \ 7 to CD9, CD9-pEF6/V5-His transfccted 293T cell

were treated with PSG 17N-Myc-I-lis and analyzed by ELISA. PSG 17N-Myc-His was

shown to bind C D9 transfected cells in a dose responsive manner with a binding plateau

at 1 0 ~lglml (Figure 1). In the cell s transfected with empty plasmid, there was only a

slight increase in background binding with increas ing PSG protein concentration. Non-

64

E 2 1

I ~ j ---<>- Empty Plasmid

= -e- mCD9 Transfected '" 0

" ;; ~ u §

05 i .D ~

0 ~

.D

O~: : « : 0 5 10 15 20

~g/111 1 PSG 17N

Figure I. Binding of PSGI7N to murine CD9 transfected 293T cells. I-IEK 293T cel ls were transfectcd with CD9-pEF6/V5-H is or empty plasmid. Forty-eight hours after transfection, increasing concentrations of PSG 17N-Myc-I-li s were added for I hour at room temperature. Binding of PSG \7 was detected after treatment wi th H RP conjugated anti-mye Ab fo llowed by addit ion ofTMB/pcroxidase substrate. The data is expressed as mean absorbance ± S.E. Each data point represents five identica l wel\s from three independent experiments

16

§ 1.4

1.2 '" 0 .,. " u 0.8 u ~ 0.6 .D " 0 0.4 ~

.D « 0.2

0

0 2 468 Ilglml Ab Concentration

---<>- PSG 17 + Match(-d Isolypc Ab

__ PSG 17 + Anti-C09 Ab

10 12

Figure 2. Anti-C09 antibodies inhibit PSG.7 binding. I-IEK 293T cells transfectcd with murine CD9-pEF/V5-l-lis were treated with increasing concentrations of an an li­CD9 Ab or an isotype matched control Ab. PSG 17N-Myc-His was added at a fixcd concentration of 5 Ilg/ml to each welL Unbound PSG 17 was washed from the well and I-IRP conjugated anti-myc Ab was added . Each data point represents four wells and is expressed as mean absorbance ± S.E.

65

specific binding of the anti-myc Ab to CD9 and control transfectcd cells was similar to

binding of 5 ~lglml PSG l 7N to control cells (data not shown). This indicates that the

binding reported in the control ce ll s reflects non-specific binding of the secondary

antibody. Rosettlllg experiments further confinned binding of PSG 17N-Myc-His to

CD9. CD9-pEF6/V5-His transfected cells treated with PSG 17N-M yc- His and anti -myc

Ab and rabbit ant i-mouse Ig imt11 unobeads showed a rosetting pattern. This pattern was

not observed in the cont rol di shes, lacking ei ther PSG I 7N treatment , anti -myc Ab or

expression of CD9 (data not shown).

To demonstrate specific binding between PSG 17N and murine CD9, competition

ELiSAs were perfonned in HEK 293T ce ll s transfecred with empty vector or C09-

pEF6/V5-l-lis (F igure 2). Forty-eight hours after transfection, anti-CD9 antibody or a

matched isotype control were added followed by PSG 17N-Myc-His. The add ition of 10

~lglmt anti-CD9 Ab reduced PSG 17N binding to background levels. PSG 17 binding to

CD9 was also analyzed by flow cytometry (Table I; Figure 3). HEK 293T cells were

transfected with C09 -pEF6/V5-His or empty vector. Treatment with PSG 17N or other

control proteins and a series of ant ibodies followed (the combinat ions of antibodies are

disp layed in Table I.) To determine extent of expression of C09 in the transfected cells,

the cell s were treated with biotin labeled anti-C09 Ab or a matched isotype.

Approx imately 70% of the cells stained positive for CD9 (lane 16 figure 3A). The

percentage of cells that sta ined positive for PSG 17 binding to CD9-transfected cells

ranged from 66-87 % while there was less than I % binding to the empty vector

transfected ce lls (Figure 3A and 8). Screening of the library by panning and binding

experiments had been performed on cells of the same li neage. To examine whether

66

PSG 17 binds to C09 on ce lls of a different species, we transfeeted BHK-2l cells wi th

CD9-pEF6/V5-1-lis, The result s show significant binding of the PSG 17N to C09

transfectcd BHK-2 1 cells (~97%) compared to controls «5%) suggesting that C09

expression may be suffic ien t for PSG 17N binding in any cel l type (Table 2)(Figure 4).

67

Table 1. Treatments used in Flow Cytometric Analysis of PSG binding to CD9 in HEK 293T Transfected Cell Type IS

{ Incubat ion 2nd Incubation 3rd Incubation 1. 293T (empty vector) 2. 293T (empty vec tor) J. 293T (C09)

x' x X X X X

X X X

4. 293T (empty vector) PSG! 7N ·1 Anti-Mye G",M I gG-Bio~

5. 293T (C09) PSG 17N AllIi-Myc G",M IgG-Bio

6. 293T (empty vector) X 7. 293T (CD9) X 8. 293T (cmpty vcclor) PSG6N 9. 293T (C09) PSG6N 10. 293T (cmpty vector) PSG 18N II. 293T (C09) PSG 18N 12. 293T (empty vector) PSG 19N 13. 293T (C09) PSG 19N 14. 293T (empty vector) Anti-CD9-Biotin 15. 293T (empty vector) Rat IgG2 a, Biotin 16. 293T (CD9) Anti-CD9-Biotin 17. 293T (CD9) Rat IgG-k Biotin • x: W",h bnne, , S,,,,p-FITC; Strcpt" , td",· Fl TC (0.5 It~) , All PSG r r",6n, ar~ M ~~' Hi, bb< kd: 10 ILg w~rc added h' ~ad, tub<

' (j . ~ll gG_ Hio: G"a[ An"· ~ l uu;c IgG, •• -A,o,in \..ar..: lo;d IO. 5 ,'g)

' y. No [!>Cuba[ ;"" ,[e r

A.

'" , +---.c-~~~~-_. __ ... I ... I.~ I ' 3 ~ 5 (, 7 8 Q IU II 12 13

••

'"

y !

Y Y Y

1~1~ 1 " 17

Sample _

B.

y y y y

<

".

•

O· , " ~ " • ' .

" .

Detection X

Strcp-F ITC ~ Strep-FITC

Strcp-FITC

St rcp-FITC

• • • ~lUOl"'f!S«!IlY Im.",oi,,' ''>

Figure 3. FACS analysis of PSG17N-Myc-His binding to CD9 transfected 293T cells. (A) 293T cel ls transfected with empty plasmid or murine CD9-pEF6/V5-His were treated with PSG 17N-Myc-His, ant i-myc antibody, biotin conjugated goat anti-mouse JgG2aK

and FITC conjugated-streptavidin. The percent binding was detennined by flow cytometry. Each sample withou t a bar represents an average binding of < I %. See Table I for treatment protoco l. (B) Al ignment of flow cytometry reports for binding of PSG 17N to empty plasmid transfected 293T cells (#4)(dotted line) and binding of PSG 17N to the C09 transfeeled ce ll s (#5)(sol id line).

68

..

Table 2. T reatment protocol for PSG binding to CD9 in BHK-21 cells by flow cytomctry Transfected Cell Type I. BHK21 (empty vcctor) 2. BHK21(cmpty vector) 3. BHK21 (CD9)

lSI Incubation 2nd Incubation x' x X X X X

3rd Incubation Detection X X X Strep-FITC2

X Strep-F ITC

4. BHK21(cmpty vector) PSGI 7N " Anti-Mye G ", M IgG-Bio4 St rep- FITC

5. BHK21(CD9) PSG I7N Anti-Myc G ", M IgG-Bio Strep-F ITC

6. BHK21 (CD9) PSG18N 7. BHK21(cmpt y vector) Anti-C D9-Biotin y -'" Y 8. BHK21(crnpty vcclOr) Rat l gG2 "~ Biol in Y Y 9. BHK21(CD9) Anti-CD9-Biotin Y Y !1~O,;. !!B";H~K"2\;1~(l.C",D,,9,,) ___ -,R>J,!!,,-,I,,,,gG 2"""B";"0<,,;,,,, _--'y'---_ _ __ ---Ly _______ _ ' X: WJsh buffer

Slrep-f I te S l rtpt,l\·l dlll·~ J fC \ (J ~ lI!;t ' Art I'S(i I'mlcins are t'o Jyc- tl is lab.:led: to II!:! " 'cre added hlcad l tubt

' (; ~ t'o J tgG-Bio: Gual Al\li-M()u~c jg(j !~ ·l3 ioli n Labelcd (1) .5 pg)

' Y: No secund and th ird incubal ion sleps

A.

'''' 1 " j

100

" " ; 0 00 • < • '" , ,

20

0 , •

S' ''' pl< •

, , Sam pte II '"

B.

•• " , i " , \ i \

.:! :' V "

\ \ ,

Figure 4. FACS analysis of PSG.? binding to BHK-21 cells expressing murine CD9. (A.) BHK-21 ceJls transfected with empty plasmid or murine CD9-pEF6/V5-His were treated with PSG l7N-Myc-His, an ti-myc antibody. biotin conjugated goat anti -mouse IgG2aK and FITC conjugated-streptavidin (Table 2), The percent binding was detemlined by flow cytometry. Sample #9 demonstrates expression ofCD9. (8.) The dotted line represents empty plasmid transfected 293T ce ll treated with PSG 17N (#4) and the so lid line represents binding of PSG l7N to the CD9 transfected cells (#5).

69

Binding of Alkaline plw!iphafase-PSG 17 (A p.PSG 17) to CD9

The previously presented binding experiments utilized a truncated N-domain of

PSG 17. We examined binding of the full length PSG 17 to murine CD9 using a

recombinant AP- PSG 17 fusion protein. HEK 293T cells wcrc transfected in 6 well poly­

L-Iys ine coated dishes and treated with increasing concentrations of AP-PSG 17 or an AP

control protein (Figure 5). Full length PSG 17 binding to CD9 transfectcd cells was

concentration dependent while binding of the AP-control protein remained at baseline

leve ls.

Murine PSGIS a"d Murine PSGJ9 do 11 0 1 himl to CD9

We have previously shown that PSG 18 induces cytokine secretion in RAW cells

( 10) and preliminary data from our lab showed that PSGI9 has the same effect suggest ing

that receptors for PSG 18 and PSG 19 are present on these ce ll s. To determine if these

murine PSGs utilize the same binding site as PSG 17. we tested binding of these proteins

to CD9 transfected 293T cells. Binding of PSG 18 and PSG 19 to CD9 was investigated

using AP-fus ion prote ins. flow cytometry, or ELISA. In the AP-binding assays, CD9-

pEF6/V5-His transfected cells were treated with two different concentrations of AP-

PSG 17, AP-PSG 18 or the AP-control protein. Binding of PSG 18N was equiva lent to

tha~ of the AP control protein (Figure 6). PSG I8N and PSG19N also did not show

significant binding 10 CD9 transfect ed ce ll s by flow cytometry (F igure 3A and 4A) and

by ELISA (data not shown). Our results indicate that these two murine PSGs do not bind

to the tetraspanin C09.

70

7000

6000 -O-- ,\p

5000 ___ AP-PSG I 7

" , ·1000 , , = ;;.. JOOO 0

2000

woO

0

0

1>1\1 Tr~at{'{]

Figure 5. Full length PSG)7 binds to C D9 transfected cells. 293T ce ll s transfected with C D9-pEF6/V5-His were treated with increasing concentrat ions of AP-PSG 17 or the AP-control protein. Each experiment was done in triplicate and the data points represen t tbe mean concentrat ion bound. The standard error was less than 25 pM for all samples.

3500

3000

~500

" ;: 2(1(10 ~ :e 1500 ~

lOOO

500

AP AI'-r se l 7 AP-r SC I IIN

Figure 6. PSG18N and PSGI9N do not bind to C D9. 293T cell s were transiently transfected wi th murine CD9-pEF6/V5-His. AP-fusion proteins were added for 90 minutes at room temperature. The concentration of bound prolein was detennined as previously described. Each bar represents the mean of at least 3 data points. Standard error was less than 15 pM for all samples.

71

Discussion

In thi s report. we identified the first binding partner for a murine PSG family

member as we ll as the firs t natural ligand for a member oftbe terraspallin superfamil y.

Using va rious methods, we showed that murine PSG 17 binds to the tetraspanin famil y

member C0 9. The binding between PSG 17 and CD9 is specific. as it ean be inhibited

with increas ing concentra tions of anti-CD9 antibody. Two other murine PSGs. PSG 18

and PSG 19 do not associate with thi s receptor. This is very intrigu ing because murine

PSG 18 and PSG 19 have conserved biological functions with PSG 17. however these

effects may not be triggered through identica l receptors.

Sequence ana lysis ofC09 suggests that like other tetraspanins. there arc four

hydrophobic transmembrane domains. These four domains create two ex tracellu lar

loops, a small and long one, with two short intracytoplasmic tail s at the amino and

carboxyl tcmlini (20). In the plasma membrane, C0 9 associates with other tetraspanin

family members including CD53, CD63, CD81. CD82, and CD 15 1 (2 1, 22). C D9 has

al so been immunopreeipitated with several p- integrins (23, 24). Other cell surface

molecules present in tetraspanin complexes indude HLA-OR (22) and MHC Class 11

glycoproteins (25) . PSG 17 could signal induct ion of cytokines by means of variety of

interactions wi th any of these molecules or others ( 13). To date, the complexes

associated with CD9 in murine macrophages have not been detennined.

I'!uman and murine CD9 di splay 89% homology at the amino acid level (26),

which together with the conserved fun ct ion of human and murine PSGs suggests that at

least one of the human PSGs could also uti li ze th is binding partner. We have tested

recombinant human PSG I, PSG6 and PSG I I proteins for binding to 293T ce ll s over~

72

express ing human C09. None of these three glycoproteins associated with this

tetraspanin. Co-immunopreclpilalion expcriments with a peptide deri ved from the N­

domain of human PSG II identified a 46-kOa prolein from TI-IP-I ce ll membranes as a

puta tive binding mati f on these cells for PSG I I (6). In terestingly, some human

telraspanins have simililr apparent molecular weigh ts as the binding molecu le ident ified

in the human monocytic cell line. Presently. we arc testing whether any o f these three

human PSGs bind to other tetraspanin fami ly members. however. it is possib le thai

human PSGs have evolved to utili ze a differen t receptor entirely.

Two groups have independent ly generated CD9 knockout mice and di scovered

that these mice have severe fertility defects (27). C09 expression on ooeytes is utili zed

for sperm-egg fusion resulting in reduced fertility for C0 9-1- femal es. We expec t that

treatment of peritoneal macrophages from C09 knockout mice with PSG 17 will not

result in the up-regulation of Il- IO. 1l-6 and TGFPI , however we may al so find that

other receptor(s) may al so bind PSG 17 and result in cytokine induction. C09 -1- mice

can produce some viab le fetuses. indicating that CD9 is important for fertilizat ion but is

not essentia l for pregnancy success. This observation is consistent with the fact that

although murine PSGs seem essential during pregnancy. they have evolved to use

different receptors and no single PSG may be criti ca l. Also we anticipate that PSG­

mediated cytokine secretion would be more crucia l in malings of mice with different

genetic backgrounds.

J-Iuman C09 is expressed on the surface of extravi llous trophoblasts (23). With

its associalion to the illlcgrin (15, human C09 is suspected to play an important role in

regulation of trophoblast invasion in humans. The possib le role of C 09 in invasion of the

73

murine placenta rema ins to be investi gated. The assoc iat ion of PSG 17 and CD9 has

linked the goa ls of better defining PSG immuno-regulatory con tro l of the innate immune

system during pregnancy and determining the signaling events following binding to a

murine tetraspan in .

References

I. H. Bohn, Arcli. Gl'Iwko/. 210, 440-57 ( 1971). 2. T. M. Lin. S. I'. Halbert, W. N. Speliaey,J Clil/. II/vesi. 54.576-82 (1974). 3. R. M. S il ver, K. D. Heybomc, K. K. Leslie, P/accllla 14, 5S3-9 (1993) ; L. L.

Arno ld el al., Am. 1. Reprod. Immlillol. 41 , 174-82 (1999). 4. I-I. Bolm, E. Weinmann. Arch. GlIl1akol. 217, 209- 18 (1974). 5. S. Snyder el III., AJRI 45. 205-216 (2001 ). 6. K. J. Rutherfurd, J. Y. Chou, B. C. Mansfield, Mol. E'ldocrillol. 9, 1297-305

( 1995 ) 7. F. Ruden, A. M. Saunders , S. Rebstock, J. A. Thompson, W. Zimmennann,

Mamm Genome 3, 262-73 ( 1992). 8. G. Rettenberger, W. Zimmennann, C. Klett, U. Zechner, H. Hameister.

Chromosome Res 3, 473-8 (1995). 9. B. Kromer et aI., E lf/". J. Biochem. 242, 280-287 (1996). 10. J. Wessell s el aI., Ell/". 1. Immunol. 30, 1830- 1840 (2000). II. R. Waterhouse, 1. Wesse ll s, K. Whi le, G. Dveksler" (200 1). 12. K. Nakamura, T. Mitamura, T. TaJ(~hashi , T. Koba yashi , E. Mekada,) Bioi Chem

275, 18284-90 (2000); C. Lagaudriere-Gesbert el al., Celllll1l11tuto/182 , 105- 12 (1997); F. Berditehevski, E. Odintsova, J Cell Bioi 146, 477-92 (1999).

13. H. T. Maecker, S. C. Todd, S. Levy,Faseb J II , 428-42 ( 1997). 14. V. Serru, P. Dessen, C. Boucheix, E. Rubinstein , Biochim Biophys Acta 1478,

159-63 (2000). 15. B. Seed, A. Aruffo , Proc Natl Acod Sci USA 84,3365-9 (1987). 16. G. Kaplan el aI., The EMBOJountIl/ 15, 4282-4296 (1996). 17. A. Aruffo, B. Seed, Proc Natl Acad Sci USA 84, 8573-7 (1987). 18. G. Kaplan, A. Levy, V. R. Raeanie lio, J Virol63, 43-51 (1989). 19. J. G. Flanagan, H. J. Cheng, Meiliods Ellzymol 327 , 198-2 10 (2000). 20. F. Lanza el at., ) BioI Chem 266, 10638-45 ( 199 1 ); C. Boucheix el at., ) BioI

Cltem 266,1 17-22 ( 1991). 2 1. G. Horvath el aI., J Bioi Cltem 273, 30537-43 (1998). 22 . E. Rub instein el al., Eu,-) 111111/11/10/26, 2657-65 ( 1996). 23 . T. Hirano e l al., Mol Hum Reprod 5, 162-7 (1999). 24. B. Baudoux, D. Castanares-Zapatero, M. Leclercq-Smekens, N. Bema, Y.

Poumay. EliI') Cell Bio! 79, 41 -5 1 (2000); K. R. Park et al., Mol Hum Reprod 6, 252-7 (2000); E. Rubinstein, V. Poindessous-Jazat. F. Le Naour, M. Billard, C. Boucheix, Eur) Imflluno/ 27. 19 19-27 (1997).

25. P. Angeli sova, L Hilgert . V. Horej si . Immunogenetics 39, 249-56 (1994).

74

26. E. Rubinstein, M. Billard , S. Plaisance, M. Prenant , C. BOllcheix, Throlllb Res 7 1, 377-83 ( 1993).

27. K. Miyado el al., Scien ce 287, 321-4 (2000); F. Le Naour el al., Leukemia II , 1290-7 (1997).

75

IV. Discussion

PSG Induction of Cytokines in Pregnancy

IL-IO and Macrop/wges

Macrophages are requi red for implantat ion in mice [ 19 1] and macrophages are

prevalent in the uterus during pregnancy [141, 192]. We showed that the N-domain of

PSG 17 induced secretion of Il-I 0 in SA L131c thioglycollate-induced peritoneal

macrophages and in RA W 264.7 ce ll s, a murine macrophage cell line. A minimum

concentration of 10 Ilglm l was necessary to up-regulate Il - IO production. and thc

cytokine response was dose-dependent in RAW 264.7 ce ll s. Uterine macrophages havc

been characterized as " immunosuppressive" during pregnancy [193]. Macrophage­

secreted Il-I 0 can down-regulate inflammatory cytokines produced in activated human

and murine mononuclear cells such as IF N-y, TNFu, granulocyte macrophage colony

stimula.ing fae.or (GM-CSF). IL-I and IL- 12 [194-196]. IL-IO also inhibilS .he

express ion of IL-8 and G-CSF at both the mRNA and protein levels [1 97]. In addition,

IL-I O regulates the express ion level of cell surface molecules required for antigen

presentation on cenain classes of human monocytes [198]. Even in the presence of IL-4

or IFN-y, IL- I 0 decreases the expression of MHC class II molecules on LPS-acti vated

monocytcs inhibiting antigen presenta tion and thus inhibi ting T cell activation [199}. In

the mouse model of RSA, inflammatory cytokines such as IFNy and TNFa are associated

with spontaneous abonioll [ 107]. Exogenolls IL- I 0 can suppress these cytokines and

prevent abonion in these mice [102} . The ab ility of PSGs to induce IL-IO synthesis in

macrophages suggests that PSGs may playa ro le in down-regulating inflammatory

76

immune responses to produce the TI-I2-like immune environment that is thought to

prevent fetal rejection.

IL-IO Expression in Trophoblast.'>

Placental trophoblast cells const itut ively produce I L-I 0 ill vitro and in vivo [9 8,

99, 132}. Production of IL- I 0 by these cells may al so contribute to the TH 2 cytokine

profile observed dUring pregnancy [200). Trophoblast cells treated with exogenous I L-l 0

showed a decrease in matrix metalloproteinase (MMP)-9 secretion and invasion ill vilro

[20 I]. MM Ps are a family of ex trace ll ular matrix degrading enzymes with functional

roles in ti ssue remodeling. These experiments suggested that trophoblastic lL-IO Illay

regulate trophoblast invasion of the uterine wall. Converse ly, lL- l 0 may have a dual role

in regulating trophoblast invasion, because decreased production oflL-1 a is associated

with preeclampsia, wh ich is characterized by shallow invasion of the placenta into the

uterine wa ll [202,203]. Reduced serum concentrations of PSGs are also associated with

preeclampsia in women [7]. Therefore, PSGs may not only influence the immune system

response during pregnancy, but may help control placental invasion through their ability

to regulate IL-I a production. As described later, PSGs may also use autocrine pathways

to influence the abil ity of trophobiasis 10 secrete lL-l O.

Interleukin 6 and Pregllallcy

IL-6 is a multifunctional cytokine with diverse biological activit ies including

regulation of hematopoiesis. and proliferation and function ofB and T cells, macrophages

and hepatocytes. Besides lymphocytes. macrophages/monocytes. microglia, astrocytes

77

and mast cell s. IL·6 is produced by many non· immune-rclated cel ls and organs. including

trophoblasts and endothe lial ce ll s [204]. There is a significant increase in IL-6

expression by uterine epithelial cell s and macrophages at the beginning of pregnancy

[205,206]. In mice. IL-6 bioactivity is greatest on days 5 and 6 of pregnancy [90]. III

vivo experiments with lL-6 knockout mice demonst rated that thi s cytokine is important

for immune response regulation. In the absence of IL-6. an over zea lous secret ion of

inflammatory cytokilles occurs after treatment with LI)S or bacte ri al infection [207, 208].

PSG induction of IL-6 in macrophages could moderate inflammatory immune responses,

which are believed to be hannfu l to the fetus.

The Importallce of TGF-pl during Pregnancy

The importance ofTGF-p in the induction and maintenance of normal pregnancy

is suggested by TGF·p-knockout mice. These animals have severe. spontaneous. mult i­

organ, auto immune disease associated with autoant ibody production. which proves fatal

in early life [209, 21 0]. TGF-~ has multiple immunosuppress ive effects at the cellular

level. inhibiting B cells, CD4+ and CDR" T ce ll s, macrophages and natural killer ce ll s

[2 11]. We showed that treatment with on ly I IJglml PSG 17N induced production o f

TGFP I in RAW 264.7 cells. Albeit with different kinetics. PSG 17 treatment of peritoneal

macrophage also resu lted in an up-regulation ofTGF~I' PSG induction ofTGFP I could

further cont ribute to the immunoregulatory effect s induced by IL- I O. TGFp can also

downregulate MHC class II expression on macrophages thereby suppressing the abi lity of

APCs to activate T cells.

78

Prostaglalldin £] and Pregnancy

Prostaglandins are eicosanoids or li pid mediators. Prostaglandins arc potent

paracrine homlones that are essential fo r a variety of functions affecting both fetal and

maternal ti ssues during pregnancy [212. 213) . PGE2 can upregulate the levels of both lL-

10 and IL-6 produced by macrophages [214], which made it an ideal potential mediator

for induct ion of both oflhese cytokines. Furthermore, PG E2 inhibits production ofTH 1

cytok ines and has been suggested to playa role in supporting TI--I2 immune responses

[2 15-2 17). The product ion of IL- l0 has a negative feedback effect on PG E2 secretion by

macrophages, which may explain why the levels of PGE2 in RA W cell supernatants were

not elevated at six and twenty-four hours post-treatment. The ability of PSG 17 to induce

this eicosanoid in macrophages provides an additional mechanism for suppression of

immune responses that are potentiall y hannflll to the fetus .

CD9- PSG I 7 Interaction

To iso late the putative PSG 17 receptor, a RA W 264.7 ce ll cDNA expression

library was constructed. Our binding experiments indicated that RAW cells expressed

approx imately 1770 PSG 17 binding sites per cell with a Ko of2.2 X 10- 11 M. Although

the receptor number was lower than in peritoneal macrophages, the init ial binding data

we obtained with RAW 264.7 cell s was more reproducible than that obta ined fcom

primary macrophages. Therefore, due to both the reproducibility of binding and the ease

of preparation of good quality mRNA needed for the library, we decided to use the cell

li ne as a source of RNA.

79

To identify the idea l cell line to use for the transfection of the library, we looked

at several different parameters. First and most important ly, the ce ll line had to be

PSG 17· receptor negative and had to have low background binding to PSG 17 or the anti­

myc antibody used to coat the plates. Second, the ce lls would have to be transfected with

high effic iency and idea ll y wou ld maintain the plasmid episomall y for easier recovery.

Other important features were the mortality rate after treatment with the selection reagen t

and the time required tor se lection. In other words, we were interested in being ab le to

select in a fast and reli ab le way, only the ce ll s that were transfected with library DNA.

We determined that human 293 cel ls did not bind PSG 17 us ing an AP-binding

assay (described in paper # 1). The AP control protein and AP-PSG 17 showed identica l

binding to these ce lls. Background binding was lower in 293 cell s when compared 10

other putative receptor negative cell s such as NIl-I 3T3 cell s. To ensure the highest level

oftranslection effi ciency, we tested a nu~~er of trans feet ion reagents using a GFP·

reporter plasmid (generously supplied by Dr. Sil vio Gutkind ). This plasmid used the

same promoter, elongation factor I -a, to drive expression oflhe GFP eDNA, as present

in the vector the RAW cell library was in. The transfeclion efficiency in EBNA 293 cells

was subjectively scored using a fluorescent microscope 24 and 48 hours post transfect iotl.

The highest percentage oftransfected cells was obtained using Lipofcctamine 2000 (Life

Technologies), followed closely by GenePORTER (Gene Therapy Systems, San Diego.

CAl with between 85-95% positi ves. Calcium phosphate lransfection was by far the least

effective method (40-50% positives). For the establishment of stably transfected ce ll s.

293 EBNA cell s were treated with 0.5 Ilglml puromycin, the antibiotic resistance gene

present in the library plasmid. Puromycin was added 48 hours posHransfcct ion and its

80

concentration was increased to I pglml one week post·transfection. As demonstrated

using the GFP control. thi s regimen resulted in effective killing orthe nOlHransfected

cells by three days wi th no toxicity for the transfected ones.

Once the cci l line was se lected and the diffcrent parameters were worked out. we

proceeded with screening of the li brary as described in paper #2. Sequencing of clones

that conrelTcd PSG 17 binding on transfected 293T ce ll s revealed that the receptor for

PSG 17 was the tetraspanin C09. We continllcd the C09-PSG 17 binding by ELISA.

now cytomctry. AP binding assays and ill sit ll reselling. We analyzed whether

transfcction of C09 resu lted in PSG 17 binding in a PSG 17- receptor negative cel l from a

different species, BHK·21 ce ll s. In addition. we attempted to transfect NIH3T3 cells, but

the trans feet ion efficiency, for the GFP- control veclor was significantly lower in these

ce ll s so they were not used.

In parallel with the li brary screening. we designed experiments aimed al the

characterizat ion of the size ofthc putative reccptor(s). Although Mansfield and

coworkers had identified a single molccule binding to human PSG II in THP-I ce ll s. a

human promonocytic cell line [51 }. we also wanted to verify that the murine PSG 17

receptor was composed of a single polypeptide chain. For this purpose, RA W cell

membranes were labeled with NHS-LC Biotin (Pierce). To isolate the receptor(s) , the

biot inylated membranes prep was passed through a PSG 17N-Myc-His-nickel-NTA

column (Qiagen). After several washes, the bound proteins were separated from the

beads by boiling in 6X SDS loading buffer. The proteins were separated on a gradient

SOS-PAGE gel. prior to Westem blot ana lys is. The membrane proteins on the

nitrocellulose membranes were treated with streplavidin-horseradish peroxidase (I-IRP)

81

and detected on film by chemilumi nescence. To identify proteins which were non­

specifically bound. the assay was repeated several times using Ni-NTA beads coated with

PSG l 7N or psp-amis, a hi stidine- labeled control protein from the Pseudomonas plilrida

(supplied by Luva Grinberg, USUHS). Three different "putative" receptors were

identified in thi s manner at 58,42. and 27 kDa. (The other proteins may have nOIl­

specifically bound to the beads or were componems of PSG receptor containing

"membrane rafts" that sometimes foml in the presence of mild detergents . Detergent

strengths were varied with each successive column in attempts to prevent thi s

phenomenon from occurring.) To confinn PSG 17N binding to the "putative" receptor

protein on the blot, PSG 17N-Myc-His was added to the membrane, and anti-myc Ab was

used to detect binding. Unfortunately, the anti-myc Ab would al so detect contaminat ing

PSG 17N-Myc-His on the blot obta ined from the affinity column upon boiling. Due to

the similar size ofCD9 and PSG l7N-Myc-His. thi s approach would not have allowed us

to di stinguish between presence of the 24 kDa CD9 and the presence of the PSG! 7N­

Myc-His in the gradient gel.

It wil l be important to detennine whether CD9 is a component of the

signal pathway leading to PSG 17 mediated cytokine secretion in macrophages. To this

end. we are currently breeding CD9+/- heterozygotes (generously provided by Dr. Claude

Boucheix) to produce CD9 knockout mice. We expect that PSG 17 treatment of

peritoneal mac rophages from C09 -/- mice with PSG 17 will not result in the up­

regulation of ll- IO, IL-6 and TGF~l. Treatment ofC09·/- maerophages with lPS wi ll

be used as a positive control for eytokine secretion . In the event that PSG 17 induces

82

cytokine production in the absence ofCD9 expression, further screening the eDNA

library for add itional receptors will be done.

Sites a/Expression o/"CD9

The discovery of the PSG 17 -C09 interaction has increased the importance of

identifying sHcs of cxpression of thi s tctraspanin, particularl y wi th rega rd to cel ls of the

immune system. Ori tan i el (II has already reported C09 expression on a variety of

hematopoieti c cell types, Two-dimensional FACS anal ysis of pc rip her at lymphocytes

from mouse spleen identified C09 expression on Band T ce lls [177]. Murine CD9 was

also fo und on peripheral blood platelets. neutrophils and mature myeloid cells [1 77J.

C09 was identified on eosinophils (1 70) and on acute non-lymphoid leukemias [175]. In

addition, C09 expression was confimled in RAW 264.7 cell s (data not shown) and in

cultured dendritic ce ll s (Dr. Jesus Colino, USU HS, personal communication).

CD9 and T cel/s

T lymphocytes are a key component of the immune response . C09 is present on

activatcd T cells. Therefore, to widen the investigat ion of PSG effects to include the

adapti ve immune system. future functional studies of PSG 17 should also include T cel ls.

Fujiwara and co lleagues demonstrated proliferation of naive murine T cells in the

presence of anti-C0 3 and anti-C09 an tibodies [218], Within 72 hours, apoptos is of anti­

C03 and anti -C09 Ab treated ce ll s occurred while cells treated with anti-C03 and anti ­

C028 Ab remained viab le [1 80). The C09- CD3- induced activation ofT cell s only

transient ly up-regulated IL-2 production ( less than 24 hours), whereas anti-C028

83

stimulated cells constitutive ly secreted lL~2. The an li-C D9/anti -C03 Ab activated ce ll s

were rescued from apoptos is when exogenous IL-2 was added 10 Ihe cultures [180].

Contrary to these resu lts, when the human T cell line lurkat is acti vated with anti-CD3

Ab and then treated with anti-C09 Ab these cells secrete IL-2 and no apoplOsis occurs

[18 1]. The role of CD9 in T cell proliferation shou ld be fUl1her investigaled using its

natura l li gand.

PSG 17 binding 10 C09 on T cell s may resul t in the modulation of cytok ine

induction. To determine the etfects of PSG 17 on T cell s, mixed lym phocytes and/or

purified T cell s from CD9 knockout mice will be stimulated with anti -C03 and/or anti­

C028 in the presence of PSG 17, and cytokine secrelion and cell proliferation wi ll be

examined. Preliminary data on mixed lym phocytes from BA LB/c mice treated with

PSG 17N revealed an inhibition of proliferation of acti vated ce ll s in the presence of

PSG 17. A reduction in proli fera tion was observed in cc ll s treated with PSG 17 and CD3

and/or C028 implying that murine PSGs could repress activat ion of these immune ce ll s.

IL- IO suppresses expression of the co-stimulatory ligands for C028, CD80 and C0 86,

thereby abrogating interactions that are essential for T ce ll act ivation [Ding, 1993 #530;

[2 19]. IL- IO also induces alloantigenic-specific unresponsiveness in human C08+ T

cel ls activated in the presence of profess ional APC [220] . The ability of PSGs to induce

macrophage IL- IO secretion may prevent CD8+ T cell s from proliferation after antigen

presentation. Studies using purified T cell s wi ll demonstrate whether the effects of

PSG 17 on proliferation are dependent on I L-I 0 secretion by APCs. Suppression of

maternal T ce ll act ivation has been reponed during pregnancy, however the mechanisms

that inhibit proliferation to prevent fetal react ivity arc speculat ive. We propose that

84

PSG 17 is inducing I L- I 0 secretion in macrophages and potentially a lso in T cells through

interaction with C09. This increase in I L-I 0 production could be a contributing factor 10

regulat ion of T cell acti vat ion during pregnancy. IL-IO inhibits I L-2 and TN Fa

production in anti-C03 activated T cells in the absence of APCs [198.22 11 . The abili ty

of PSG 17 to induce I L-\ 0 secretion in act ivated T cel ls should be investigated.

CD9 E,pressioll ill NOIl-Hemalopoielic Tissues

Murine C09 expression has also been found in non-hematopoietic tissues.

Northem blot analys is has indicated thai C0 9 is transcribed in the heart, kidney, lung.

spleen. bra in. nerves, thymus and li ver in mice [ 17 1, 177. 222]. Western blots con finned

protein express ion in hea rt. kidney, lung, intestine. muscle, spleen and brain [1 77}.

Reports of ti ssue express ion in hOlllogenates of heart and liver should be caut iously

interpreted since these preparations may be contaminated with blood cells. The

significance of CD9 expression in stromal tissues is uncerta in. Treatment of stromal or

myeloid ce ll s with anti-C09 antibodies enhances adhesion of these cells and results in

inhibition of migration, suggesti ng that signaling through C0 9 augments the adhes ive

properties of both cell types. Recent pub lications correlated the expression ofC09 to the

ability of certain tumors to metastasize [223]. Interesting ly, uterine invasion by the

placenta has been equated to metastatic invasion of cancer cells. although placental

invasion is generall y a controlled process. The poss ible ro le ofCD9 in invasion of the

murine placenta should be investigated. The expression of C09 by the anchoring

placenta l structures suggests that th is telraspanin could regulate endometrial invasion.

C09 is expressed in invasive trophoblast ce ll s in humans. IfC 0 9 is also expressed on

85

simi lar cells in mice , PSG 17 induct ion of IL-I 0 through interaction with CD9 could

suppress uterine invasion by trophoblast ce ll s.

Signaling Mechanisms

rite PSG / 7N sigl/aling cascade

PGE~ production by macrophages within 2 hours of PSG 17 trealmen t suggested

that synthesis of this eicosanoid is a direct result of PSG 17 binding to ils receptor. In

contrast, upregu lat ion of IL- I 0 and IL-6 mRNA occurred later, suggesti ng that de novo

protein synthes is was necessary to upregu late mRNA produc tion of these cylOkines after

treatment with PSG 17. The necess ity of de novo protein synthes is was confirmed by co­

treatment with cycloheximide, which prevented increased production of both IL- IO and

IL-6 mRNA. Induction of both IL-6 and IL- I 0 protein secretion requi red a mini mal dose

of 10 J.l.g/ml PSGI7N. suggesting that the,same receptor is utili zed 10 induce both

cylokines. As previously stated, PGE2 has been shown 10 induce both IL-IO and IL-6 in

activated murine macrophages [2 14]. PGE;: biosynthesis is di rect ly mediated by the

cycloxygenase enzyme, COX-2. Interestingly, cycloxygenase-2 deficient mice are

plagued with reproductive fa il ures [224]. We showed that PSG I7 induced PGE2 within 2

hours of treatment at the same minimal concentration that induced I L- l 0 and lL-6.

Whether the induction of PGE! by PSG 17 is mediated by Cox-2 will be invest igated with

pharmacological inhi bitors of this enzyme, such as indomethacin, NS-398 or SC-5823S.

Induction of PGE2 could provide PSGs with a pivota l means of regulat ing act ivated

macrophages and system ic cytokine production. In add ition, PGE2 can increase

macrophage levels of cycl ic-AM P, which in tum activates 1 L-IO and IL-6 expresss ion via

86

transcription factors such as CREB and API. Therefore. PGE2 regulation in response to

PSG 17 may be a critical part of the signal transduction pathway leading from receptor

activation to cyfokine secretion.

CD9 Signaling Events

If PSG 17 induction of cytokine secret ion through CD9 binding is confirmed. we

intend to examine the signaling events invo lved in this pathway. All of the tetraspanlll

fami ly members contain four putat ive transmembrane domains, two extracellular loops

and two short cytoplasmic tail s. Thc short cytoplasmic tails oftetraspanins suggest that

they may signal through other associated proteins. Identifying the C0 9 associated

proteins requires gentle extraction to prevent di sruption of the association of tetraspanin

molecules with other tetraspanins and with integrins [1 56, 163]. Murine CD9 has been

shown to associate with other tetraspanins including C063. C081, and CD82 [ 161 ].

CD9, CD63, CD8 l and CD82 have been found in complex with HLA-DR and VLA

integrins as con finned by immunoprecipitation and immunoblotting oftransfected

fibroblasts and B lymphoma cell lines [1 6 1). CD9 also associates with various 131

integrin molecules including a Jj3 l, 0-4131 , a sj3 l, a6J31 and allbJ33 [225-228]. A number or

tetraspanins, including C09, specifically associate with a 55 kDa type II

phosphatidylinositol-4-kinase in lymphoid and carcinoma cells [229]. CD9 has also been

linked to G proteins but only in platelets [230, 231]. The variety of possible signaling

cascades that these associations provide suggest mechanims by which tetraspanins can be

linked to their proposed functions of cell motilit y, growth and cell differentiation. There

87

is currently no infonnat ion regarding associat ions of C0 9 with Olher tetraspanins, cell

surface or intrace llular molecules ill macrophages.

With the almost ubiquitous expression of CD9 on murine hematopoietic ce ll s. one

question Ihat arises is "How might the CD9-PSG 17 interaction affect stromal cells or

other hematopoieti c deri vatives such as megakaryocytes and plate lets?" The presence or

absence of signaling components and binding partners within these cell s could potentially

dctennine the biologica l effect of the receptor-ligand interaction. There is a consensus in

the literature that C09 and other tctraspan in protein complexes may be cel l type and

acti vat ion slate dependent [154]. suggesting that signaling cascades through this receptor

arc cell-specific. Therefore, unt il the components of these complexes arc determined the

effects of PSG 17 on various C09-expressing ce ll s can not be predicted.

\-\lorking toward a Comprehesive Mouse Model of PSG Function

The ;1/ vitro biological effects of murine PSGs arc being investigated for the

purpose of deve loping a viab le animal model for the study of human PSG function. Our

results indicate that despite the structural differences between murine and human PSGs,

they both have similar biological functions;n v;tro. The complex ity and size of the

murine and human PSG fami lies have made the ident ification of thei r function very

difficult Although all human PSG cONAs have been eloned and sequenced, much

further research needs 10 be done with the murine PSG family. Reagents that can identify

individual PSG family members may be impossib le to obtain due to thei r similarity at the

amino acid level. In the next sections. I will describe the too ls our laboratory is

deve loping towards the goal of creating a mouse model of PSG funct ion.

88

Generatioll of a Murine Anti-PSG Anlibody

Anli-murine PSG antibodies will be useful in characte ri zing binding and function

of thesc proteins. Pre liminary att empts to produce PSG ISN-GST antibodies were not

successful. Unfo rtunately, the antibodies developed were not PSG IBN-spec ific and

bound to a variety of other denatu red proteins on Westem blots [232]. Another attempt

to produce anti -murine PSG anti body is current ly being pursued using recombinant

PSG 17N-Myc-His prote ins. Ant i-muri ne PSG antibodies can be used to characterize the

molecular weight of secreted PSGs ill vivo. and to detenninc the concentration of murine

PSGs in serum throughout gestation. As was observed with human PSGs, we anticipate

that murine PSG anitbodies will not be specific for ind iv idual PSGs.

Knockouts and Transgenes

Notwithstanding the differences in receptor specificity, some murine PSGs have

overl apping biological functions. Therefore, we would expect that thc genetic defic iency

ora single PSG receptor would not be incompatible with successful pregnancy. Indeed,

C09 knockout females can produce a few viable fetuses [188, 189]. Macrophages from

CD9-/- mice will also be useful in confirming that PSG 18 and PSG 19 use different

receptors from PSG 17. Because these murine PSGs did not bind to C0 9 in in vitro

experiments, we expect them to induce cytoki nes in macrophages from CD9-/- mice.

We have already developed a PSG 17N transgenic mouse, which will allow us to

study in vivo PSG 17 function outside of the contex t of pregnancy. We created the

transgenic construct by ex tract ing the PSG 17N-Myc-His cDNA from the pcONA3. 1

vector and insert ing it into pEAK 10 (Egde Biosystems). Wi thout anti -murine PSG

89

antibodies. the myc-his tag was necessa ry to confinn prote in express ion in vivo.

Expression of PSG 17N-Mye-I-lis in the pEA K I 0 vector is dri ven by the elongation factor

I alpha promoter (ELF I a). ELF I a is a strong promoter for expression il/ vitro and ill

vi l'o. and therefore. we anticipate that PSG 17N will be expressed in most. if not all.

tissues. Ten pronuclei wcre injected and two founders were iden tified. The transgenic

animals were init iall y sc reened for the presence of the insert by PCR. To confi rm PSG 17

DNA was transcri bed in these animals. RNA was ex tracted from severa l tissues and

analyzed by RT-PCR. I'SG I7N was transcribed in several organ tissues including lung.

hean , kidney. illlestines and spleen of progeny from one of the two transgenic founders.

Several studies arc planned afler we confi ml the expression of PSG 17 prote in. PSG 17 !

transgenic mice will be crossed wi th mice that are suscept ible to autoimmune di seases,

such as experi mental acute encephaliti s (EAE). EAE is used as a model for mult iple

scleros is, and pregnancy has been shown to affect the severi ty of thi s and other

autoi mmune di sorders. This system could be used to detcnnine whether the consti tuti ve

production of PSG 17 leads to remiss ion or at least repress ion of symptoms including

paralys is. The effects of high levels of PSG 17 during the course of an immune response

to several well -characteri zed pathogens would also be interes ting to examine. Serum

levels of both TH 1 and TH2 cytokines should be measured in the transgenic animals, and

compared to control mice that had or had not been challenged with pathogens.

Idel11!DJi llg additional hl/ lIIGlI and murine PSG receptors

We are currently pursuing the ident ity of the receptor(s) for human PSGs. Human

and murine CD9 di splay 89% homology at the amino acid level (1 871. which suggests

90

that at least one of the human PSGs might also ut ili ze this binding partner. We have

tested recombinant human PSG I, PSG6 and PSG II proteins for binding to 293T ce ll s

over-expressing human CD9, and none of these PSGs assoc iated with this tetraspanin.

We are systematicall y testing for interactions between available human tctraspanin clones

and these three human PSGs. The majority of amino acid differences between human

and murine CD9 tctraspanins are located in the larger of tile two extracellular loops

[1871, and the expression pattern ofCD9 also differs slightl y between the two species

[233). Therefore, it is possib le that human PSGs have evolved to utili ze a different

receptor mechanism ent irely. Regardless of whether human PSGs bind to a member of

the tetraspanm superfam il y, the similarities between immune regulatory funct ions of

PSGs in mice and humans justify the usc of a mouse model of PSG funct ion. The

identification ofa tetraspanin family member as a receptor for human PSGs would only

strengthen thi s model.

Investigation of additional receptors linking murine PSGs with immunoregulatory

function wili help us to define the role of individua l murine PSGs during pregnancy. To

identify other murine PSG binding sites, a graduate student in our lab is currently

screening the RA W cell expression library for the murine PSG 19 receptor. Treatment of

RA W ce lls with PSG 19N-Myc-His also increased secretion of IL- I 0, IL-6 and TGFP],

indicating that activation of a different PSG receptor(s) induces similar bio logical

functions in these cell s. We are also continuing to screen for binding of murine PSG 18

and PSG 19 to murine tetraspanin family members. The tetraspanin family members,

CD53 and CD81 are only expressed in speci fi c hematopoietic cells, which makes them

interest ing potenti al PSG targets [1551.

91

Murine PSG 17 binds to the tClraspani n CD9. This is the first interact ion defined

for both of these glycoproteins. The tools for determining the PSG signaling cascade as

well as the bio logica l funct ion ofa tctraspanin fam il y member, arc now at our di sposal.

We will cont inue to pursue the identifi cation of other PSG receptors. Generat ion of a

PSG transgenic mouse and ident ificat ion of the interaction between PSG 17 and C0 9 has

provided LI S with the tools necessary to design in vivo stud ies in the future. This

infonllation is vital for di ssecting the mechanisms by which PSGs induce cytok ine

mR NA and exert any other poten tial effects on the matemal immune system.

92

V. Appendix 1

PSG Nomenclature

Current literature reviews and intemet searches reveal that a variety of names

are in use for PSG genes and proteins. 1·luman PSGs are referenced as human pregnancy

- specific ~ I-glycoprotein ( PS~G or PBG), (very frequently) Scwangerschaftsprotein

(SP I ). pregnancy-associated plasma protein C (PArr-C) and trophoblasl-specific~ 1-

glycoprotein (TSG). Murine PSGs are called murine Carcinoembryonic antigens (mCeo),

and rat PSGs are referred to as rnCGM. Several proposals for new PSG nomenclature

have becn recommended [234. 235] in attempts to standardize the incongruent

identification ofPSGs in this growing field of research. To avoid confusion in this

manuscript, all references reflect the November 2, 1999 CEA nomenclature

announcement published in Experimental Cell Research 252(2) 243-249. PSG family

members are referred (0 as PSGs followed by a gene identification number. Human

PSGs genes are designated PSGI-II. Mouse PSG genes are now named PSG16-29 and

rat PSG genes follow as PSG36-40. Italics are used 10 indicate the gene and no rmal font

indicates the corresponding protcin.

93

VI. Appendix 2

Purification of PSGI7N-Myc His

PSG J 7N Myc-H is/pFastBac was transfected into S/9 ce ll s with Cellfectin (Life

Technologies) . The vinls was titcred by plaque assay according to manufacturer 's

instructions (Life Technologies). To produce prOlein, 1-li5 ce ll s were inoculated with a

multiplicity of infection of I in Ex-400 Media (Irvi ne Scientific), and the ce ll

supernatants were harves ted 72 hours post infection. Protease inhibito rs (I % aprotinin ,

1% leupcpt in , 1% phenylmcthylsulfon yl fluoride (PM SF) Sigma) were added to prevent

protein degradation. Insect cell supernatant containing recombinant secreted protein was

concentrated with an Am icon Ultrafiltration 8400 Ce ll and YM-3 ultrafilter (Milliporc).

The concentrated supernatant was dial yzed in 300 mM NaCI; 50 !TIM NaH2P04 dialysis

buffer using snakeskin dialysis tubing (MWCO 10,000 Oaltons)(Pierce). PSG 17N·Myc

His was prec ipitated from the dialyzed supernatant with Ni·NTA beads (Qiagen) in the

presence of 10mM imidazole in dial ysis buffer to decrease non·specific binding. Bound

proteins were eluted from the beads with increas ing concentrat ions of imidazo le . Thc

eluted proteins were concentrated in a Centriprep-3 (Mi llipore).

The recombinant protein was then separated by electrophoresis in a 12% SOS

PAGE gel in a mode1491 Prep cell apparatus (BioRad) and collected in test tube

fraction s as they migrated out of the gel. PSG 17N·Myc·I-lis-containing fraction s, as

detennincd by Western blot with anti·myc Ab. were pooled and c1ectroe luted in 10 mM

Tris buffer, pH 8.0 in a Sia lomed apparatus (Amika Corp. , Columbia. MO) to remove

SOS from the sample. The proteins were concentrated with a ccntriprep·3 (Mi llipore).

The final protein concentration was ascertained by comparison to bovine serum albumin

94

standards on a 4-12% Bis Tris SDS-PAGE gel (Inv itrogen) stained with GelCode blue

reagent (Pierce-Endogen) Llsing the Eagle Eye II Still Video system and Eagle Sight

version 3.1 software (Stratagene). Limulus amebocyte lysate assays (Biowhittaker, MA)

were perfonned to ensure low levels of LPS contamination. Protein samples were

deemed negative only used if endotoxin levels were less than 0.2 ng/ml which would not

result in inc reased eytokine secret ion in macrophages (Dr. Stephanie Vogel, USU HS.

persollal COIll11l1Illicatioll).

95

VII. References

I . Bolm, I-I., [Detectioll and clwrClclerization (~lpregnan(y proteins inlhe human placenta and tlleir quantitative immlillochemical determination in seraji'olll pregnant women]. Arch. Gynakol. , 1971. 210(4): p. 440-57.

2. Dimitriadou. F .. et al. , Discordant secretion o.f pregnancy specijic bela 1-g~Jlcoproteil/ and 11Ilmall chorionic gonadutropin by JUfman pre-embryos clflfllred ill vitro. Fertil StcriL 1992. 57(3): p. 63 1-6.

3. Lin, T.M .. S. P. I-I albert , and W.N. Spellacy. Measurement o.fpregnancy­associated plasma proteins Juring human geslatioll. J. C li n. In vest.. 1974 . 54(3): p. 576-82.

4. Lee, J.N .• J .G. Grudzinskas. and T. Chard . Cireulalillg levels o/pregnancy proleins in ear(jl and late pregnancy in relation to placenlal tissue cOllcelllmlion. Br J Obstet Gynaecol. 1979. 86( II ): p. 888-90.

5. Tamsen. L. , Pregnancy-specific bela l-g(v'coprotein (SP I) levels measured l~v nephe/olllell), in serum.from women with vaginal bleeding illthejirst ha(fqf pregnancy. Acta Obstet Gynecol Scand, 1984. 63(4): p. 3 11 -5.

6. MacDonald, 0.1 ., ct aI. , A prospective study o.lthree biochemical.lelOplacental lests: serum humall placental lactogen. pregnancy-spec{/ic beta l-glycoprotein. and urinGl)' estrogens. and their relationship to placental ill.'>I!lJicie1UY. Am J Obstet Gynecol , 1983. 147(4): p. 430-6.

7. Silver, R.M., K.D. Heybome, and K.K. Les lie, Pregnancy specific beta 1 glycoprotein (SP-I) ill maternal serum and amnioticjluid: pre-eclampsia, small for gestational agejetus and jeral distress. Placenta. 1993 . 14(5): p. 583-9.

8. Hertz. J.B. and P. Schultz-Larsen, Human placental lactogen. pregIlQlu.y -spec{jic beta-l-glycoprotein and alpha-jetoprotein in serum ill threatened abortion. Int J Gynaecol Obstet , 1983. 21(2): p. 111 -7.

9. Blomberg. L.A. , et aI. , Effect oj"lwl/lall pregnancy-spec{jic betal-g(vcoprotein 011

blood cell regeneration ajter bone marrow transplantafion. Proc Soc Exp Bioi Med, 1998. 217(2): p. 2 12-8.

10 . Arnold, L.L. , et al. . PreglUlIl(y-specUic glycoprOiein gene expressioll in recurrell/ aborters: a potential correlation to interleukin-IO expressioll. Am. J. Reprod. Immunol. , 1999. 41 (3): p. 174-82.

II. Wesse ll s. J., ct aI. , Pregnancy-specific g(ycoprotein 18 induces lL-10 expressioll ill murine macrophages. EUf. J. Immuno!., 2000. 30: p. 1830- 1840.

12. Snyder, S., et al. . Pregnancy-Specfjic Glycoproteins FUllction as Immullol/lodulators by Inducing Secretion oI1L-10, lL-6 alld TGF-Beta l by Hllman Monocvtes. AJR1. 200 1. 45: p. 205-2 16.

13 . Rooney. B.C., c.1-I . Home. and N. Hardman, Molecular cloning qf a cDNAIor human pregnancy-specific beta 1- g~J'coproteill:holl/ology with human carcilloemblyollic antigen alld related proteins. Gene, 1988. 71 (2 ): p. 439-49.

14. Streydio, c., ct a I. , The human pregnan(v-specUic beta l-g(vcoprotein (PS beta G) and the carcinoemb'yonic antigen (CEA}-related proteins are members oIlhe same multigenefami(v. Biochem Biophys Res Commun. 1995. 1 S4( I ): p. 130-7.

96

15. Watanabe. S. and J. Y. Chou, /solatioll and characterization oIcomplementary DNAs encoding IlIlman pregllan(.~v-spec{fic beta I-glycoprotein. J Bioi Chern, 1988. 263(4): p. 2049-54.

16. Thompson, J .A., et aI. , Molecular clolling of a gene belol/gil/g 10 Ihe carcinoemb'yonic antigen gene/ami~JI and discussion ofa domain model. Proc NaIl Acad Sci USA, 1987. 84(9): p. 2965-9.

17. Thompson, J. and W. Zimmennann, The carcilloemb'yollic ollligell gene Jami~v: structure. expression and f'1·ollltion. Tumour BioI. 1988. 9(2-3): p. 63-83.

18. Thompson. J.A .. F. Grunen. and W. Zimmermann. Carcilloembl:vonic alltigen gel/efal1/i~lI: molecular biologv and clinical penpectives. J Clin Lab Anal, 1991 . 5(5): p. 344-66.

19. Thompson, J.A .. Molecular cloning and expression ofcarcilloemb,:mllic anligell gene/all/ify members. Tumour BioI. 1995. 16( I): p. 10-6.

20. Strcydio. C .. et aI. , Strllclllre. evolution and chromosomal localization <?(the hllmall pregllallcy-spec{{ic bela I-glycoprotein gelle/amily. Genomics. 1990. 7(4): p. 66 1-2.

21. Thompson, J.. el aI. , The human pregllall91-specific g~vcoproteilJ genes are tightly linked on the long arlll of chromosome J 9 and are coordil/are~JI expressed (published errafllm appears ill Biochem Biophys Res Commllll J 990 May 16: 168(3): IJ25j. Biochem Biophys Res Commun, 1990. 167(2): p. 848-59.

22. Khan , W.N., el aI., The pregnancy·specijic glycoproteinjamily of the illlmullogiobulin sllpelfami(Y: idelllf{icatioll ofnew members Gnd estimation qf fami~vsize. Genomics, 1992. 12(4): p. 780-7.

23. Teglund , S. , et aI. , The pregllallcy-spec~lic glycoproteill (PSG) gene duster 011

human chromosome 19: fine strllclllre of the J J PSG genes and identification oj 6 new genes forming a third subgroup within the carcinoemblyonic antigen (Ct.A) fa/lli~v. Genomics, 1994. 23(3): p. 669-84.

24. Olsen, A. , et aI., Gene orgallization o/the pregnanc.:)I-~pecijic glycoprotein region 011 IlIlmall chromosome J 9: assemb~v and alla~vsis o/a 700-kb cosmid cOlltig spanning the region. Genomics. 1994. 23(3): p. 659-68.

25. Zheng. Q.X. , ct aI. , Characterization o/cDNAs of the humall pregnancy-specific beta J -glycoprotein family. a lIew subjall/if)' qf the immulloglobulin gene slIpclfamily. Biochemistry, 1990. 29( I I): p. 2845-52.

26. Hefta, S.A., et aI. , Carcilloembl),ollic antigen is anchored to membranes by covalent attachment to a glycosylphospharidylinosiro/ moiety: idemijicatioll oj the ethanolamine linkage site. Proc all Acad Sci USA. 1988. 85(13): p. 4648-52.

27. Ruden. F., W. Zimmennann, and J.A. Thompson. hllra- and illterspecies analyses of the carcilloemb,yollic amigen (C1:A) gellefamify reveal independelll ello/lttion in primates alld rodents. J Mol Evol. 1989. 29(2): p. 126-34.

28. Oikawa, S., ct aI. , A specijic heterotypic ceJl adhesion activity between members qfcarcilloembl)lonic antigen/ami/yo W272 alld NCA, is mediared by N-domaills. J Bioi Chem, 199 1. 266( 13): p. 7995-800 I.

29. Benchimol, S .. et aI. , Carcilloembryonic antigen, a hllmanlUmor marker. jUllctions as all illtercellular adhesion molecule. Cell , 1989. 57(2): p. 327-34.

30. Home, C.B ., et aI., Pregnancy specific beta I-glycoprotein--a product oflhe syncytiotrophoblast. Experientia. 1976. 32(9): p. 1197.

97

31. P10uzek, C.A., et aI., DijJeremial gelle expression ill the omuion , chorioll. alld trophoblast o/the humall placenta. Placenta, 1993. 14(3 ): p. 277-85.

32. Bischof, P., Placellfal proteins. COlltrib Gyneco l Obstet, 1984. 12: p. 1-96. 33. Wurz, H., Serum concentrations of SP I (pregnollcy-spec{{ic-beto I -glycoprotein)

in healthy. lIonpregnant individuals. and in paliellls with lIo/lIropllObfastic malignant lIeoplasms. Arch Gyncco l, 1979. 227( I): p. 1-6.

34. Shupen, W.L and W. Y. Chan, Pregnancy spec{{ic beta I-g~\lcoproteill in hUlllall inteslil/e. Mol Ce ll Biochem, 1993. 120(2): p. 159-70.

35. Borjig in. J ., ct al.. Expressiol/ of the pregnal/cy-specijic beta I -g(vcoproteill gel1es ill humalltestis. Biochem Biophys Res Commun, 1990. 166(2): p. 622-9.

36. Chan. W. Y .. et al.. PregnCln(v-specijic hera I g(H:oproleill mRNA is presel1l ill placental as well as nOli-placental rissues. Hum Rcprod, 1988.3(5): p. 677-85 .

37. ZOllbir, F .. W.N. Khan, and S. J-Iammarst rom. Carcilloemb/yullic an ligen gene .!lll1li(V members in submandibular sa fivalY gland: demonsrrarioll o/pl·eg/wl/c.:\,­specijic glycoproteins by eDNA cloning. Biochem Biophys Res Commlln , 1990. 169( 1): p. 203-16.

38. Chan, W., The Pregnllncy Speq/ic pi G~\'coproteil1 Family. Adv Contra Deliv Syst, 199 1. 7: p. 21 -52.

39. J-Iome. C.H., I.N. Reid , and G.D. Mi lne, Prognostic sigl1(jicllnce oIinappropriate prodllction o.f pregnollCY proteins by breast cancers. Lancet, 1976. 2(7980): p. 279-82.

40. Walker, R.A ., D{f]erentiatioll of human breast carcinomas: all immullohistological study o.{appropriare and inappropriate protein productioll. J Patho!. 1981. 135( I): p. 87-95.

41. Rosen, S. W. , ct al., Human fibroblasts produce "pregllan':v-spec{jic " bera-I g()lcoproteil1 il1 viTro. Am J Obstet' Gyneco l, 1979. 134(7): p. 734-8.

42. C hou, J .Y .. ProduCTioll o/pregnancy-specijic beta }-glycoprotein by hllmon placenral cells and hllmon / ibroblosts. Oncodev Bioi Med. 1983 . 4(5): p. 319-26.

43. Oikawa. S. , ct ai. , A pregnancy-specific beta I-glycoprotein. a CEA gelle/ami~v member. expressed ill a human promye1oC)'tic leukemia cell line. HL-60: structures of protein. mRNA and gelle, Biochem Biophys Res Commun, 1989. 163(2): p. 1021-31.

44. Bamett , T.R., W.d. Pick le, and lJ . Elting. CharacterizllIion o.fllvo lIew members of the pregnancy-specific beta I-glycoprotein ji:lIl1ily ii-om rhe myeloid cell line KG-} and suggestion O/IWO distinct classes o!trallscripriollllnit. Biochemistry. 1990. 29(44): p. 10213-8.

45 . I-I e ikinheimo, M., ct al., Ollcoplacental prorein SP I --a COllstitlllive and inducible late di!ferenriarioll marker o.f the humall ",yeiomOnO(Ftic lineage. Blood, 1987. 70(5): p. 1279-83 .

46. Sea rl e, F., et ai. , Serum -SPI -pregnancy-specijic-beto-glycoprotein in choriocarcinoma and other neoplosric disease. Lancet, 1978. 1(8064): p. 579-81.

47. Fagnart. O.c., et aI., Prognostic vallie 0/ concentration 0/ pregllallc.:v-spec([ie beta I -glycoprotein (SPJ) in serum o/patients with breast cancer. lnt J Cancer. 1985. 36(5): p. 54 1-4.

48. Zimmennann, W., M. Weiss. and J.A. Thompson, cDNA c/ollingdemonslrates the expression o.fpregnallc),-spec(/ic glycoprotein genes, a subgroup 0/ the

98

carcilloemb,)'ol1ic allligel! genefamity . ill/etal liver. Biochemica l and Biophysica l Research Communications, 1989. 162: p. 1197- 1209.

49. Kuroki. M .. Y. Koga, and Y. Matsuoka , Monoc/ol/al amibodies to carcilloembl),onic antigen: a !}J'stematic analysis q/allfibo(~11 spec~ficities by /Ising related normal antigens and evidence for alforypic determinal1rs 011

carcilloembl)'ollic al/tigen. J Immunol , 1984. 133(4): p. 2090-7. 50. Thompson. J. , el aI. , A polymerase-chain-reaction assay/or the spec~{ic

ident{jication q/transcl"ipts encoded by individual carcil/oemh'yolI;C antigen (C£A)-gel/e~fallli~v members. Int J Cancer, 1993.55(2): p. 3 11-9.

5 1. Rutherfu rd , K.J., J. Y. Chou. and 8.c. Mansfie ld, A motif ill PSG lis mediates binding to a receptor 01/ the sw:j(lce qfrhe promonO(I'le eelliine THP-1. Mo l. Endocrino!. , 1995. 9(10): p. 1297-305.

52. Ruos laht i, E.,lnlegril1s. J Clin Invest. 1991. 87( [): p. 1-5. 53. Ruos laht i, E. and M.D. Pierschbacher, New perspectives ill cell adhesioll: RGD

and inlegrins. Science, 1987. 238(4826): p. 49 1-7. 54 . Wu, S.M., el aI., Erpressioll o/pregnallcy-specUic beta I-g(vcoproleill gelles in

hell1atopoieticcel/s. Mol Cell Biochem, 1993. 122(2): p. 147-58. 55. Majumdar, S .. et aI. , Pregllam.:v speeUk proteills: suppression qj"ill vitro

blastogenic response to mitogen by these proteins. Int J Fertil, 1982. 27(2) : p. 66-9.

56. Watanabe. S. and J.Y. Chou, HUlllan pregl1al1cy-!}pec~fic heta I-glycoprotein: a nell' member oflhe carcilloembl),ollic anligen gene/ami/yo Biochem Biophys Res Comlllun, 1988. 152(2): p. 762-8.

57. Wu, S.M., et aI. , Effect qf"pregnancy-specijic beta I-glycoprotein 011 the developmelll of preil11plantation emblYo. 1999. 220(3): p. 169-77.

58. Ho, P.c., S.Y. Chan , and G.W. Tang, Diagnosis of early pregnancy by enzyme immunoassay qf SchwangerschaIts- protein I. Ferti! Steril, 1988. 49( I ): p. 76-80.

59. Kromer. 8., et aI. , Coordinate expression of splice variaf/ls q{the murine pregnancy-spec{jic glycoprotein (PSG) geneIamily durillg placental development. Eur. J . Biochem., 1996. 242: p. 280-287.

60. Rudert, F., et aI. , Characterization q{murille carcilloembryonic antigen gene ./il/nily members. Mamm Genome, 1992. 3(5): p. 262-73.

6 1. Snyder, S. , Human Pregllancy-Spec~{ic Glycoproteills FUllctioll as Immullomodulators ill Vitro by Inducing Secretion qf lL - / 0 and IL-6 in Human Monocytes, in Molecular alld Cellular Biology Program. 2000, Unifonned Services University of the Health Sciences: Bethesda, MD.

62. Saunders, A.M. and M.F. Seld in, A molecular genetic linkage map o/mollse chromosome 7. Genomics, 1990. 8(3): p. 525-35.

63. Finkenzeller, D. , et a!. . cea5, a st/"llcturally divergent member q{ the murine carcinoemblyollic antigen gene/amily, is exclusively expressed duril/g early placental development ill trophoblast giant cells. J Bioi Chern, 1997. 272(50): p. 3 1369-76.

64. Pijnenborg, R., W. B. Robertson, I. Bronsens, and G . Dixon, Trophoblast Invasion and the Establishment of Haemochorial Placentation ill Man and LaboratOlJ' Allimals. 1981 , London: W.B. Saunders Company Ltd. 7 1-89.

99

65. Chan, W. Y., et at., Characterizatioll of cDNA ellcodillg human pregllallcy­spec~jic bera J -glycoprotein/rom placenta alld extraplacental tissues alld their comparisoll with careinoemblyonic alltigen. Dna, 1988.7(8): p. 545-55.

66. Chen, 1-1 ., et aI., Characterization a/a major member o/rhe rat pregnaney­speedicglycoprotein/ami~v. DNA Cel l BioI , 1992. 11 (2): p. 139-48.

67. Tease LA, S.W.L., Fazleabas A T, Zavy M T, Chan W y, Presence of pregll(/luy-spee!fic PI-glycoprotein in d{fferellt species o.f"allimal. Proceeding of Ihe XVllllh Meelillg ISOBM 1989: p. 62.

68. Niemann, S.c. , et aI. , Pregnml(l'-spec~fic beta /-g~~ .. coprotein: cDNA cloning. tissue expressioll. alld species spec~ficity q( olle memher oIthe PS beta C.fallli~F. Hum Genet, 1989.82(3): p. 239-43.

69. Zhou, G.Q. and S. Hammarstrom, Pregnancy-Specijic C~vcoprofeil/ (PSG) ill Baboon (Papio hamadryas): Fallli~v Size. DOlllain Structure. alld Predictioll o(a Functional Regioll in Primate PSGs. BioI Reprod, 200 I. 64( I ): p. 90-99.

70. Towler, C.M., et aI., Plasma levels oIpregnan(v-specdic beral-glycopralein in normal pregnancy. Br J Obstet Gynaecol , 1976. 83(10): p. 775-9.

71. Balm, H. and E. Weinmann, [Immunological disruptioll o/implantarioll ill monkeys with alllihodies 10 humall pregnancy specUk beta / -g~vcoproleill (SP I) (OI I1f1or's lrans!)}. Arch . Gynakol. , 1974. 217(2): p. 209- 18.

72. Tamsen, l., S.G. Johansson, and O. Axelsson, Pregllancy-spec{/ic bela 1-glycoprotein (SP I) in senllll Fom women willi pregnancies complicated by illtrauterine growth retardation. J Perinat Med, 1983. li e I): p. 19-25.

73. Bohn, H. and E. Weinmann, {Allt~{ertiliry e.ffect q{an active immtlllizalioll of monkeys with human pregnancy-spec die beta I-glycoprotein (SP I) (allthor's lrallsl)}. Arch Gynakol , 1976. 221 (4): p. 305-12.

74. Hau, J. , ct aI. , The effect on pregnancy of intrauterine administratioll 0/ antibodies against two pregnancy-associated murine proteins: murine pregllall(jI­specdic beta J -g~Jlcoproteill and murine pregnallc:JI-associared alpha 2-glycoprotein. Biomed. Biochim. Acta, 1985. 44(7-8): p. 1255-9.

75. Home, C.H. , C.M . Towler, and G.D. Milne, Detection o/p,.egnallcyspec~jic beta I-g~vcoproteill in j(Jlmalin-jixed tissues. J e lin Pathol , 1977. 30( 1): p. 19-23.

76. Tatarinov, Y.S., D.M . Falalecva, and V.V. Kalashn ikov, Human pregnancy­spec{fic beta I-globulin and its relation 10 chorioepithelioma. Int J Cancer, 1976. 17(5): p. 626-32.

77. Les li e, K.K., et aI. , Linkage a/two human pregnancy-spec(fic beta I-glycoprotein genes: one is associated with hydatid(form mole. Proc Natl Acad Sci USA, 1990. 87( 15): p. 5822-6.

78. Seppala, M. and M. Purhonen, The use o/hCG and other pregnancy protein.~· in the diagnosis 0/ ectopic pregnancy. Cli n Obstet Gynecol , 1987.30(\): p. 148-54 .

79. Chamberlin, M.E. , K.J. Lei, and J .Y. Chou, Subtfed!'fferences in humall pregnancy-spec[fic g~ycoprOiein gene proll/Olers aliowjor differential expression. J Bioi Chern, 1994. 269(25): p. 17152-9.

80. Medawar, P.B., Some immunological and endocrinological problems raised by evolutioll a/viviparity in wrIebraIes. Soc. Exp. BioI.. Evolution Symposia. Vo l. II. 1953, New York: Academic Press. 320-338.

100

8 1. Billingham. R.E. and P. Medawar. ACfive~v acqllired folerance o.f foreign cells. Nalure, 1953. 172: p. 603-606.

82. Hamilton. M.S., Marel'l1al immune responses to ollco/eral antigens. J Reprod Immunol, 1983. 5(5): p. 249-64.

83. Nagarkatti. P .S. and O.A. Clark. III vitro activity alld in vivo correlates of alloalltigen-speciJic murille suppressor T cells induced by allogeneic pregnancy. J

Imillunoi. 1983 . 131 (2): p. 638-43. 84. Bell, S.c. and W.O. Billington, Allli:/etal a"o-alltibo{~jI in the pregllalltfemale.

Immunol Rev. 1983.75: p. 5-30. 85. Beer. A.E .. R.E. Billingham, and R.A. Hoerr. Elicitation and expression 0.(

transplantation il11l11ul1ity ill the utel'lls. Transplant Proc. 197 1. 3( I ): p. 609-1 [. 86. Hunt. l S. and J .W. Pollard. Macrophage.\· ill the uterlls and placenta. Curr Top

Microbiollmmul1OI. 1992. 181 (1): p. 39-63. 87. Vince. G .. et al., Localization o/tulllournecrosisjactor prodllction ill cells at the

matemo((etal inleljace ill human pregnamy. Cli n Exp Immunol, 1992. 88( I): p. 174-80.

88. Clark. O.A ., et al.. Characterizolion (~fmurine pregnancy decidua lrall~/o/,lIIillg

growth [aclOr bera. 1. Trallsforming growth /actor beta 2-like molecules of lI1lusLwlmolecLlIor size released in Moaclive form . Bioi Reprod. 1995. 52(6): p. 1380-8.

89. Vince, G.S., ct aI., Flo\\' L:vtometric characterisalion o.f cell poplilations ill humall pregnalll.:r decidua and isolatioll oj decidual macrophage .... J Immunol Methods, 1990. 132(2): p. 18 1-9.

90. Oe. m .. Sanford, T.R. , Wood. G.W .. Expression o.f IL-I. /L-6 and TNF-alpha ill mouse IfterllS during the peri-implanrarioll period. J. Reprod Ferti l, 1993. 97: p. 83-89.

9 1. Winkler, M., et al., Regulatioll of illterlellkill-8 .\y"lhesis in humall lower uterine segmelltJibroblasts by cytokines alld growth /actors. Obstet Gynecol, 2000. 95(4): p. 584-8.

92. Pollard. J. W .. Lymphohematopoietic cytokilles in the female reproductive tract. Curr Opin Immunol, 1991. 3(5): p. 772-7.

93. Jones, C.A., J.1 . Finlay~J ones, and P.H. Han, Type- I and type-2 cytokines in human late-gestation decidual tissue. Bioi Reprod, 1997. 57(2): p. 303-1 1.

94. Kauma, S .. et ai. , Interlellkin-I beta, Itllman leukocyte antigen HLA-DR alpha, and transjorming growlhfhclOr-beta expression in endometrium, placenta. and placental membranes. Am J Obstet Gynecol, 1990. 163(5 Pt I ): p. 1430-7.

95 . Stephanou, A. , et aI. , Ol/lOgeny o/the expression and regulation oj interleukin-6 (IL-6) and IL-I mRNAs by human trophoblast cells during differel1liatioll ill vitro. J Endocrinol, 1995. 147(3): p. 487-96.

96. Masuhiro, K .. et aI., Trophoblast-derived inredeukin-I (IL-I) stimulales the release of humall choriollic gonadotropill by activating IL-6 and IL-6-receptor system ill first trimester Illiman trophoblast.\·. J Clin Endocrino! Metab. 199 1. 72(3): p. 594-601.

97. Kamcda, T. , et aI. , Production ofinterleukin-6 by normal human trophoblast. Placenta. 1990. 11 (3): p. 205- 13.

101

98. Roth, I., et ai. , Hlllllan p/acenta l cytotrophob/asts prodllce the iml/Junosuppressive cytokine illler/eukin 10. J. Exp. Med., 1996. 184: p. 539-548.

99. Cadet, P. , et ai. , lnterleukill-IO messenger rihollucleic acid in human placenta: implications ala role jar illterleukill-I 0 inretal allografi protectioll. Am J Obstet Gyneco l, 1995. 173(1 ): p. 25-9.

100. Graham, C I-f. , et aI. , Localization o/trans/hrmillg growth R lclOr-belO Of the human/etal-maternal intelface: role il/ trophoblast growlh alld dij/erelltiation. Bioi Reprod , 1992. 46(4): p. 56 1-72.

10 1. Mitchell, M.D., M.S. Trautman, and DJ. Dudley, Cytokine networking ill Ihe placel1lO. Placenta, 1993. 14(3); p. 249-75.

102 . Mosmann. T .R. and R.L. Coffman, T/-II and TH2 cells: d!f/erellf pallems of fVlllphokine secretioll lead to d{(ferellt/illlctionol propenies. Annu Rev Immunol , 1989.7: p. 145-73.

103. Gendron, R.L. , R. Farookhi , and M.G. Baines, Resorption (~lCBAIJ x DBAI2 mOllse conceptuses in CBAIJ tlteri correlates with/clilwe ~lthe/eto-placellta! lInit to suppress natural killer cell activity . J Reprod Fenil , 1990. 89( I ): p. 277-84.

104. Chaouat, G., et aL /L-lO prevents nall/rally occurillgleta! loss ill the CBA X DBAI2 matillg combinatioll. and local defect ill /L-I O prodllction in (his abortioll­prolle combination is corrected by ill vivo injection o/lFN- r. J . Immunol., 1995. 154: p. 426 1-4268.

105. Arck, P.C, A.B. Troutt , and D .A. Clark , Soluble receptors neutralizing TNF­alpha and IL- I block stress- triggered mllrineaborlion. Am J Reprod Immunol, 1997. 37(3): p. 262-6.

106. Tezabwala, B.U .. P.M. Johnson, and R.C. Rees, Inhibition qrpregnancy viability in mice/ollowing lL-2 administration. Immunology. 1989.67: p. 11 5.

107. Chaouat, G. , Synergy of lipopolysaccharide and iI~flammalOly cytokill es in murine pregnancy: a/loimmullizalioll prevents abortioll but does Iwt affect (iIe inductioll aJpreterm de/ivery. Celilmmunol, 1994. 157(2): p. 328-40.

108. Gendron, R. L. and M.G . Baines, In/iltrating decidual natural killer cells are associated with spolltaneolls abortioll in mice. Cell Immunol, 1988. 113(2): p. 26 1-7.

109. Tangri, S., et aI. , Maternal anli-placental reactivily ill nall/ral immullologically medialedjetal reSol1Jtions. J. immunol., 1994. 152: p. 4903.

110. Tangri , S. and R. Raghupathy, Expression 0/ cYlOkines in placentas a/mice undergOing immllllO/ogica//y mediated spontaneolls /etal resorptions. Bioi Rcprod. 1993. 49(4): p. 850-6.

111 . Arck, P .c.. et a I. , Regulatioll 0/ abortion by gamma delta T cells. Am J Rcprod Immunol , 1997.37( 1): p. 87-93.

112. Arck, P.C, et aI. , Murine r cell determination o/pregnancy outcome. Cell Immunol, 1999. 196(2): p. 7 1-9.

11 3. Clark, D.A. , et aI. , Cytokine-dependent abortion ill CBA x DBAI2 mice is medialed by the procoaguiant };;/2 prothombinase. J hnmunol, 1998. 160(2); p. 545-9.

114. Duclos. AJ ., E.K. Haddad, and M.G. Baines, Presence 0/ activmed macrophages ill a lIlurine model olearly embryo loss. Am J Reprod Immunol , 1995. 33(5); p. 354-66.

102

115. Haddad, EX., A.1. Duclos, and M.G. Baines, Early emblYo loss is associated lIIith local production oj nitric oxide by decidual monolluclear cells. J. Exp. Med .. 1995. 182 : p. 1143-11 5 1.

I 16. Stern, J .J. and c.B. COli lam, Current stollls oJimmunologic reClirrenl pregnancy loss. Curr Opin Obstet Gynecol , 1993. 5(2): p. 252-9.

11 7. I-li ll , J. A. and OJ. Anderson, 1II/I/I1/1/Ologicalmechallisllls in reCI/ITem spOlllaneOIlS abonion. Arch Imn1l1l1ol Ther Exp, 1990. 38( 1-2): p. 111-9.

118. Albcmlan. E .. Ill , in "Early PregJ/(lJu:\' Loss: Mechanisms and Treatment". R.W.B.a.F . Sharp. Editor. 1988. RCOG: London. p. 9.

119. Marz i. M., et al.. Clwracteri:atioll a/type I and type 1 (vtokil/e production profile ill physiologic alld h/lll/OI/ pregnanc:,'. Clin. Exp. Immunol.. 1996. 106: p. 127-133.

120. Hill . J .A .. T-lieiper I-type immullity to lrophobfast: evidellce/or a nell' illlll//IIlOfogicaf mechallism/vr reClirrenl abonion in women. H lim Reprod. 1995. 10 Suppl 2: p. 114-20.

121. Raghupath y, R .. et aL, MMemal Th 1- and Thl-type reactivity to pfacenlal antigens ill 1I0l'mal humall pregl1allcy and unexplained recurrent spolllanl!ous aborlions. Ccilimmunol , 1999. 196(2): p. 122-30.

122. Maruyama. T .. et aI. , Ffow-(:vtometric analysis ofilllllllme cell poplila/iollS ill h'lman deciduaFom variOllS types ojjirst-Irimester pregnancy. Hum Immunol. 1992. 34(3): p. 2 12-8.

123. PiccinnL M.-P. , et aI. , De/ective production ofbolh leukemia inhibitoryIactor and type 2 T-hefper cytokilles by deciduaf T ceffs in unexplained reCllrrenl aborlions. Nature Mcd., 1998. 4(9): p. 1020- 1024.

124. Lea, R.G., et al.. A subset o/patients willt recurrent spOlllaneOIlS abortion is dejicient ill trans/ormillg growth Jactor beta-l-producil1g "suppressor celts" ill lIlerille lissue Ileal' tile pfacental attachmellt site. Am J Reprod Immunol, 1995. 34( I): p. 52-64.

125. Yui , J., el aI., Cytotoxicity oIllimollr necrosis Jactor-alpha and gamma illtel/eroll against primm)' human placelliaf Iropilobfasis. Placen ta, 1994. 15: p. 8 19·835 .

126. Raghupath y, R., Th I-type immllll iry is il/compa/fbfe wilh slIccessjiti pregnancy. Immunol Today, 1997. 18(10): p. 478-82.

127. Wegmann, TG., et aI. , Bidirectiollal cytokine interactions i" tile malernaf-j(!lQf relationship: is sllccess/ul pregnancy a T,,2 phenomenon? Immunol. Today. 1993. 14(7): p. 353 -356.

128. Oc lassus, S. , C( aI. , Differelltial cylOkille expressioll ill malemal blood alld placenta dllring IIl11rine gestation. J Immunol, 1994. 152(5): p. 2411-20.

129. Li n, 1-1 ., et aI. , Synthesis oJT helper 2-lype cylOkilles at the matemal-jetaf illtelface. J Immunol, 1993. 15 1(9): p. 4562-73.

130. Chaouat. G. , ct aI. , Localization oflhe TIl2 cytokines IL-3 , IL-4, IL- IO at Ihe JelOmaternal illte/face during hUlllan and murine pregnancy and lack of requiremelli /or FaslFas ligand inleractioll jor a success/uf allogeneic pregnancy. Am J Rep rod Immunol. 1999. 42(1): p. 1- 13.

13 1. Bennett , W.A .. et aI. , First-trimester humall chorionic villi express both immlilloregufalory and inj1ammalory cylokines: a role Jor illlerieukill-IO iI/

103

regulating the cytokille network of pregllam:y . Am J Rcprod Immunol , 1999. 41 (1): p. 70-8.

132. Hanna, N. , et aI. , Gestational Age.Dependellt Expressioll oIIL·! 0 and Its Receptor ill HUlllan Placental Tissues and Isolated Cylolrophoblasts. J Immuno!. 2000. 164(1 1): p. 572 1-5728.

133. Jones, W.R., Autoimmune disease alld pregnancy . Aust N Z J Obstet Gynaccal , 1994.34(3): p. 251 -8.

134. Buyan, J.P. , The effects o.{pregnallcy all auloimll/llne diseases. J Leukoc Bioi , 1998. 63(3): p. 281-7.

135. Da Silva, J.A. and T.D. Spector, The role o.fpregl/al/(l' iI/the course and aetiologv oIrheumatoid arthritis. Cli n Rhcumatol. 1992. 11 (2): p. 189·94.

136. Vamer, M. W., A/lraimll/llne disorders and pregnancy. Semin Perinatol, 199 1. 15(3): p. 238-50.

137. Moonnann, A.M. , et al., Malaria and pregl1(1//(:V: placental cylokinc expressioll and its relatiollship to illtrallterille growth reran/arioll. J. Int'. Dis., 1999. 180: p. 1987-1993.

138. Chaouat, G., et a\. , 1lIIlIIlIl1o-endocrine interactiolls ill early pregnall(Y HUIll

Reprod , 1995. 10 Suppl 2: p. 55-9. 139. Piccinni, M.P. and S. Rornagnani , Regulation (~ffeta' allografj survival by a

hormone-coli/roiled Til 1- and TlI2-type cylokilles. Immunol Res, 1996. 15(2): p. 14 1-50.

140. Elcnkov, U ., 1. Hoffman, and R.L Wilder, Does differential neuroendocrine colltrol oIcytokine production govern the expression of autoimmune diseases in pregnancy and the postpartum period? Mol Mcd Today, 1997. 3(9): p. 379-83.

14\. Hunt, J .S. and S.A. Robertson, Uterine macrophages and environmental programming/or pregnancy stlcce.~s. J Reprod Immunol , 1996. 32( I) : p. 1-25.

142. Kovats, S., ct aI., A class I amigen, HLA-G. expressed in humanlrophoblasts. Science, 1990.248(4952): p. 220-3.

143. Picc inni , M.P., et aL, Progesterone/avol's /he development o/Iwlllan T helper cells producing Th2-type cytokines and promotes both IL-4 prodllction and membrane CD30 expression in established Til I cell clones. J Immunol , 1995. 155( 1): p. 128-33.

144. Munn, D.H ., et al. , Prevellfiol1 0/ Allogeneic Fetal Rejection by Tlyptophan Catabolism. Science, 1998. 281 : p. 1191-11 93.

145 . Kamimura, S. , et aI., Localization and developmental change Qlindoleamine 2.3-dioxygenase aClivi/y in fhe human placen/a. Acta Med Okayama, 199 1. 45(3): p. 135-9.

146. Kudo, Y. and c.A. Boyd, Humall placen/al indoleamine 2.3-dioxygenase: cellular localization and characterization Qf an enzyme preventing/etal rejectioll. Biochim Biophys Acta, 2000. 1500( 1): p. 119-24.

147. Hwu, P. , et aI. , Indoleamille 2.3-dioxygenase productioll by hUlllan dendritic cells results ill the inliibition ofT cell prohferation. J Immunol, 2000. 164(7): p. 3596-9.

148. Tatsumi, K" et al. , Inductioll OfllJlptophan 2.3-dioxygenase in the mouse endometrium during implantation. Biochem Biophys Res Commun. 2000. 274( I): p. 166-70.

104

149. Thellin. 0., et al. , Tolerance to the/oeto·placelltal 'graft': ten ways to support a child/or /lille lIIollths [III Process Citation]. Curr Opin ImmunoL 2000. 12(6): p. 731-7.

150. Morgan, B.P. and C.H. Holmes, Imlllllilology o.(repmdllclioll: pmtecring the placenta. CUrT BioI, 2000. to( I 0): p. R38 1·3.

151. Xu, c., et al.. A critical rolefor ml/rille complement regl/lalOr cn:v in .fetoll/atema! tolerance [w:e cOllllllents]. Science, 2000. 287(5452): p. 498-501.

152. Wright, M.D. , K.J. Henkle , and G.F. Mitche ll , All immllllogenic Mr 23.000 integral membrane proreill o.fSchislOsollla mallsoni worms rhat c/ose~I' resembles a hlllllall tlll/lOr·associated antigen. J Immunol. 1990. 144(8): p. 3 195-200.

153. Oren. R., et al.. TAPA · I. the targel of all (Jlltipro/ijeralil'C:! alllihody. defines a lIell" f amily (~(tralls/JI emhral1e proteills. Mol Cell BioI. 1990. 10(8): p. 4007·15.

154. Maecker. H .T., S.c. Todd, and S. Levy, The tetraspallill s llper/i:lIJli~\' : lIIoleclllar .lacili/arors. Faseb J, 1997. 11 (6): p. 428-42.

155. Wrigh t, M.D. and M.G. Tomlinson, The ills and Ollts (~rthe transmembrane 4 s lIpq!amily . Immunol Today. 1994. 15(12): p. 588-94.

156. Scrru, V., et aI. , Seqllence alld expressioll (~rsel'en nell' tetraspans. Biochim Biophys Acta, 2000. 1478(1): p. 159-63 .

157. Tachibana. 1. , el aI. , NAG-2. a novel fl"{JI1smembralle·4 slIpelfamily (TM4SF) protein that complexes with integrins alld olher TM4SF proteins. J Bioi Chern , 1997 . 272(46): p. 29181-9.

15S. Cook. G .A., et al. , The tefrCIspanin CD9 il!/luences fh e adhesion. spreading. alld pericellular./ibronectin mafrix assembly 0.( Ch il/ese hamster oVOIY cells 011 human plasmaJibronectin. Exp Cell Res. 1999. 251 (2): p. 356-7 1.

159. Shi, W ., el al., The fetraspanin CD9 associates with transmembrane TGF·alpha and regulares TGF-alpha-induced ECF receptor activalion and cell prolf(erarion. J Ce ll Bioi, 2000. 148(3): p. 59 1-602.

160. Berditchevski, F. and E. Odintsova, Characterization ofil//egrin-tetraspanin adhesion complexes: role 0.( retrCIspanins ill illtegrill signaling. J Cell BioI, 1999. 146(2): p. 477-92 .

161. Rubinstein. E .. et a l., CD9. CD63. CD8 I. and CD82 are COmpOllel11S ofa s lI/face tetraspan nerwork connected to HLA-DR and VLA integrins. Eur J Immunol , 1996.26(11): p. 2657-65.

162. Siupsky, J.R. , et aI. , Analysis o/CD9. CD32 and p67 signalling: lise of degranulated plalelels indicates direct involvement ojCD9 and p67 ill integri" activation. Br J Haemato1, 1997.96(2): p. 275-86.

163. Horvath. G .. et al. . CDI9 is linked to the integrin-associated tetraspans CD9. CD81. alldCD82. J Bioi Chern , 1998. 273(46): p. 30537-43.

164. Mannion. B.A., et al. , Transmembrane·4 sllpel1ami~V proteins CD81 (TAPA-I). CD82. CD63. and CD53 spectfically associated with inlegril1 alpha 4 be/a I (CD49dICD29). J Irnrnu nol , 1996. 157(5): p. 2039-47.

165. Boucheix. c., et aI., Molecular cloning o.lthe CD9 allligen. A new/amily of cell surface proteins. J Bioi Chern, 199 1. 266( I): p. 117-22 .

166. Sincock, P.M. , G. Mayrhofer, and L.K. Ashman, Localization o/tlre transmembrane 4 slipeliamily (TM4SF) member PETA·3 (CD 151) in 1I01"ma/

105

hUlllall tissues: comparisoll wilh CD9. CD63. alld alpha5beta I ill/egrill. J Histochem Cytochem. 1997. 45(4): p. 515-25.

167. Lanza. F., et al.. cDNA dOlling and expression of pla/ele/ p24/CD9. Evidellce for a lIew/ami(v of multiple membrane-spanning prOleills. J Bioi C hern . 1991. 266(16): p. 10638-45.

168. Higashihara. M. , et a l. . Pur~fica/ioll alld partial characterization oICD9 antigen oIllIIIIIall platelets. FEBS LeI!. 1990. 264(2): p. 270-4.

169. Doh-ura. K .. et al.. Enhallced CD9 expressioll in the //louse alld human brains iI!j(xted with transmissible spollg{jorm encephalopathies. J Neuropathol Exp Neuro l. 2000. 59(9): p. 774-85.

170. Fcmvik. E., e l al.. !lIIrac.'ellulal" alld sill/ace distrilm/ioll oICD9 ill 111/1//011

ensillopiliis. Apmis. 1995. I 03( I 0): p. 699-706. 171. Kaprieli an. Z .• el al. . CD9, all/(~ior plmelet cel! slII/i.n·e glycopl"Otein. is a ROCA

allfigell alld is expressed in the nervOlls .\ystelll . J Ncurosci. 1995. 15( 1 PI 2): p. 562-73 .

172. Nakamura . Y .. R. Iwamoto, and E. Mekada, Expressioll and distribuliol1 a/CD9 ill myelin of the central and peripheralllel"l'olls sy.Hems. Am J Pathol. 1996. 149(2): p. 575-83.

173. Kagawa. H., et aI. , Expressioll Ofpl"Olhrombillase activity and CD9 antigen Oil the slll/ace a/small vesicles ji-um stilllulated human endothelial cells. Thromb Res, 1995. 80(6): p. 451-60.

174. Zeleznik-Le. N.J. and R.S. Metzgar, Expressioll ofCD9 ollligclI ollllorllla/ activa/ed humall B cells. Ccli lmmunol. 1989. 123( 1): p. 70-82.

175. Ferrero, D .. et al.. CD9 antigen on aC LIte Iloll -(vmphoid leukemia cells: preferelltial expressioll by promyeloc.ytic (M3) subtype. Leuk Res, 1991. 15(6): p. 457-6 1.

176. Hemler, M.E .. B.A. Mannion. and F. Berd itchevski, Associatioll ofTM4SF proteills with illlegrins: relevance to cancer. Biochim Biophys Acta, 1996. 1287(2-3 ): p. 67-7 1.

177. Oritani , K., et aI. , Amibody ligation ofCD9 mod(lies productioll o.fmyeloid cells ill IOllg- term cullllres. Blood, 1996. 87(6): p. 2252-6 1.

178. Aoyama, K., et aI. , Stromal cell CD9 regulates d(/Jerelltiatiol1 of hematopoietic slem/pl'ogellilor cells. Blood. 1999. 93(8): p. 2586-94.

179. Oritani, K., et al. . Stromal cell CD9 tJnd the d!Uerelitiatioll o.fhemalOpoielic stem/progellitor cells. Leuk Lymphoma, 2000. 38( 1-2): p. 147-52.

180. Tai. X.G .. et aI. , CD9-mediated costimularioll of TCR-triggered naive T cells leads to activation followed b.v apoptosis. J Immunol, 1997. 159(8): p. 3799-807.

18 1. Lagaudriere-Gesbert, C. et aI. , Functional analYSis 0.11011 1" tetraspalls. CD9. CD53. CD8!. and CD82. suggests a commOIl role ill costimliialion, cell adhesion. and migration: on~v CD911pregliiates HB-£GF activity. Cell immunol , 1997. 182(2): p. 105-12.

182. lkeyama. S. , et al.. Suppression of cellll/otility alld metastasis by tralls/eaion with humall mOlilitJ!-relared protein (MRP-lICD9) DNA. J Exp Med. 1993. 177(5): p. 123 1-7.

183. Shaw. A.R .• et al.. Ectopic expressioll a/humall alldfelille CD9 ill a hllman B cell lillc cOl!lers bera 1 integrin-dependelll motility on./ibrollectill and lam ilTill

106

substrmes und enhanced tyrosine phosphwvlatioll. J BioI Chem, 1995. 270(41): p.24092-9.

184. Hirano, T. , et a I. , CD9 is illvolved ill invasion q/"llllman trophoblast-like choriocarcilloma cell IiI/e. 8eWo cel/s. Mol Hum Rcprod, 1999. 5(2): p. 168-74.

185. Pijncnbarg. R., ct al.. Placenral bed spiral arreries in rhe hypertensive disorders o.f"pregllallcy. Br J Obstcl Gynaecol, 199 1. 98(7): p. 648-55.

186. Korhonen, M., el al.. Distribution of the alpha I-alpha 6 illregril1 suhullits ill humall developing and ram placema. Lab Invest, 1991. 65(3): p. 347-56.

187. Rubinstein. E .. el al.. Molecillar clOllillg of rite II/ouse equivalellt of"CD9 alltigell. Thromb Res. 1993.7 1(5): p. 377-S3.

188. Le Naour. F. , et aI. , Severely redllcedje/llalefertifity ill CD9-de./iciellt /IIice. Science. 2000. 287(545 1): p. 319-21.

189. Miyado. K., et aI., Requirement o/CD9 011 the egg plasma memhrane/or jimilizmiol1. Science, 2000. 287(545 1): p. 321-4.

190. Kaji , K .. et aI. , The gametefilsioll p/"Ocess is defective in eggs 0.f"Cd9-dejicient mice. Nat Genet. 2000. 24(3): p. 279-82.

191. Clark, D.A., P.c. Arck, and G. Chaauat, Why did your/llother reject you? Immunogenetic determinallts qf"the respollse to el1viJ"OIImellwl selective pressure expressed a! the IIterille level. Am. J. Reprod. Immunol.. 1999. 41 ( I): p. 5-22.

192. Brandon, J.M .• Leucocyte distribution ill the uterus during the preimplantatioll period a/pregnancy and phagocyte recruitment to s ites 0/ blastocyst attachmellt iI/mice. J Reprod Ferti !. 1993. 98(2): p. 567-76.

193 . Hunt. J .S., L S. Manning, and G. W. Wood, Macrophages ill murine lIferus are immunosuppressive. Ceil immunol, 1984. 85(2): p. 499-510.

194. Wang, P., et aI., IL- I 0 inhibits transcriptioll oIcytokine genes in humall peripheral blood mononuclear cells. J. Immuno!.. 1994. 153(2): p. 8 11 -6.

195. Bogdan, c., Y. Vodovotz, and C. Nathan, Macrophage deaerivation by inrerlellkill 10. J. Exp. Med., 199 1. 174(6): p. 1549-55 .

196. Fiorentino, D.F. , el al. . IL-I 0 inhibits cytokine production by activated macrophages. J Immunol , 1991. 147( I I): p. 38 15-22.

197. de Waal Malefyt, R., et aI., Interlcukill /0 (fL-IO) inhihits cYfOkinesynthesis by human mOl1ocytes: all aUfOregulatOlY role o./"IL-IO pJ"Oduced by monocyte.\·. J. Exp. Mcd. , 199 1. 174(5): p. 1209-20.

198. de Waal Malefyt, R. , et a I. , Illterieukin 10 (lL-IO) and viral IL-IO strollgly reduce aJltigen-spec{/ic human T cell proliferation by diminishing the antigen-presenting capaciry o/monocytes via downregulatiol1 a/class II major hislOcomparibiJity complex expressioll. J Exp Med, 1991. 174(4): p. 915-24.

199. Hart, P.H., C.A. Jones, and J.1 . Finlay-Jones, MOllocytes cultured ill cytokine­defined environments d(fferfrom freshly isolated monocytes ill their responses 10 /L-4 alld /L-IO. J Leukoc Bioi, 1995. 57(6): p. 909-IS.

200. Bennett, W.A.. et aI. , Expression and productioll Q/"interleukin-IO by human trophoblast: relationship to pregnancy immUllOlOlerance. Earl y Pregnancy. 1997. 3(3): p. 190-S.

20 I. Roth , I. and S.1 . Fisher, IL-IO is an autocrine illhibitor oflwmall placental cytotrophoblast MMP- 9 productioll and invasion. Dev. BiaL, 1999. 205(1): p. 194-204.

107

202. Hennessy, A., ct aI. , A deficiellcy oJp/acellfol IL-1O ill preeclampsia. J Immunol. 1999. 163(6): p. 3491-5 .

203. Tay lor. R. N., Review: ilJllllllllohiology C!fpreeclampsia. Am J Reprod Immunol , 1997. 37( I ): p. 79-86.

204. Papanicolaou. D.A. , et a\., The pathophysiologic roles oIilltel'lellkill-6 il1 hlllllQI1 disease. Ann Intem Med, 1998. 128(2): p. 127-37.

205. Robertson. S.A .. G. Mayrhofer, and R.F. Seamark. Uterille epitheliol cells ~:vllthesi=e gral1l1 /oc:vte-macrop/wge COIOI1.I·- stil1l lllat illg jactor and illterlellkin-6 ill pregnant (Illd 1I0llpregllal1l mice. Bioi Rcprod. 1992. 46(6): p. 1069-79.

206. Liang, L. , ct aI. , Regulatioll C!{illterieukil1-6 and il11erleukill- 1 beta gene expressioll in rhe mouse decidlll//1/. J Reprod hnmunol. 1996. 30( I); p. 29-52.

207. Kopf. M., ct al.. Impaired immlllle alld acute-phase responses ill illlerleukin-6-dejiciellt mice. NalUrc, 1994. 368(6469); p. 339-42.

208. Xing, Z .. CI aI. , 1L-6 is all (lilt iilljla/llmafOl:v (ytokille required/or COllt rollillg loca! or systemic {lClIfe i,!/lallllllarory responses. J. Clin . Invest. , 1998. 101 : p. 311-320.

209. Chris t. M .. et aI. , Immune (~vsregulatioll ill TGF-beta I-dejicientmice. J ImmunoI, 1994. 153(5): p. 1936-46.

2 10. LCllc ri o, J .J., ct ai. , Maternal rescue o/transJorming growth /actor-bera I Ill/ II lIIice. Science. 1994. 264(5167): p. 1936-8.

2 11. Lcltcri o, J .J . and A.S. Roberts, Regulation o/immulle responses by TGF-hew. Annu Rev Immunol , 1998. 16: p. 137-6 1.

2 12. Yang, Z. M., ct aI. , Pote1llial sites l?f prostaglandin actiolls ill the periimp/alltation mouse lIIer/ls: differential expression and regulmioll C!lprosfaglulldil1 receptor gelles. Bioi Rcprod. 1997. 56(2): p. 368-79.

2 13. Grcystokc. A.P., el aI. , Tran.~(er wul metabolism oj prosraglal1d;n E(2)i1l the dllal peljilsed human placenta. Placenta. 2000. 21 (1): p. 109- 14.

2 14. Williams, J.A., C H. Pontzer, and E. Shacler, Regulation of macrophage illlerieukin-6 (IL- 6) and IL-IO expression by prostaglandin £ 2: the role olp38 mitogen-activated protein kinase. J Interfe ron Cytokinc Res, 2000. 20(3); p. 29 1-8.

215. Phipps. R.P .. S.I-1. Stein, and R.L. Roper, A new view l?(prostaglandin £ regularion oj rhe immune response. Immunol Today, 1991. 12( I 0): p. 349-52.

216. Ke ll y, R. W. and 1-1.0. Critch ley, A T-helper-2 bias in decidua: the prostaglandin cOlllrihlltioll of the macrophage and trop/lOblast. J Rcprod ImmunoL 1997. 33(3); p. 18 1-7.

2 17. Kuroda , E .. el al.. Sensitivity differellce to the suppressive ejfect (:?Iprostaglalldill £2 among mOllse strains: a possible mechanism 10 polarize Th2 type response ill BALB!c lIIice. J Il11mul1ol. 2000. 164(5): p. 2386-95.

2 18. Tai, X.G .. et aI. , A role/or CD9mo!ecliles ill T cell actillation. J Ex p Med, 1996. 184(2): p. 753-8.

219. Kubin , M. , M. Kamoun, and G. Trinchieri. f1llerleukin 12 ~:\ll/e,.gizes wilh B71CD28 illteraction in inducing ejJicielll proliferation and cYlOkine production o/llIIlIIall T cells. J Exp Med. 1994. 180( I): p. 2 11 -22.

220. Groux, 1-1 .. el aI. , Illhibifory and stimulatory 4Jects oflL-IO 011 human CD8+ T cells. J Immuno!. 1998. 160(7): p. 3188-93.

108

221. Taga, K. and G. Tosato, IL-I() illhibits human T cell pro/{Ieratioll and /L-2 production. J Immunol, 1992. 148(4): p. 1143-8.

222. Schmidt, c., et al.. CD9 o/mollse brain is implicated ill lIew'it f! outgrowth and cell migraliol1 in vilro and is associated with the alpha 6/bera I inleg,.ill alld the neural adhesiol1 lIIo/ecule L I. J Neurosei Res, 1996. 43( I ): p. 12-3 1.

223. Boueheix, e.G. H.T.D .. c. Jasmin. and E. Rubinstein., Tetraspanills[{lld Malig(1II110( 1'. Exp. Rev. Mol. Med., 200 1. www­

ermm.cbcu.cam.ac.u klO I 002381 h.htm. 224. Lim, H .. ct al.. Mulliplelemale l'epl'oductive/ailul'es ill (ycloo.\ygellase 2-

deJiciellllllice. Cell . 1997. 91 (2): p. 197-208. 225. Berdirchevski. r ., M.M. Zulter. and M.E. Hemler. Clwracteri=ation ol'lIovel

complexes all Ihe cell slIl:!(lCe between illtegrills alld proteills 1\'ith 4 fl'm/smembran£! domaills (TM4 proreius). Mol Bio i Cc lL 1996. 7(2): p. 193-207.

226. Hi rano, T. , et al.. CD9 is expressed ill extra viI/OilS Il'oplioblasts ill association with il/tegrill alpha3 alld illtegl'ill alpha5. Mol Hum Rcprod. 1999. 5(2): p. 162-7.

227. Rubinste in, E. , el aI. , CD9 al/tigen is all accessOI:I' subullit of the VLA illiegrill complexes. Eur J Immunol, 1994. 24(12): p. 3005-13.

228. Nakamura. K. , R. Iwamoto, and E. Mckada, Membrane-anchored heparill­binding ECF-like gro\l'rh/aclOr (HB-£CF) and diphtheria toxilll'eceptor­associated protein (DRAP27)/CD9 jonn (J complex with illtegrin alpha 3 beta I at cell-cell contaci siles. J Ce ll BioI. 1995. 129(6): p. 169 1-705.

229. Yauch, R. L. and M.E. l'lemler, Spec~fic interacrions amollg rrallsll/embrane 4 supel.1tllllily (TM4SF) proteins alld pllOspllOil/osifide 4-kinase. Biochem J, 2000. 351 PI 3: p. 629-37.

230. Seehafer. J.o. and A.R. Shaw, Evidence that the signal-initiating membrane proteill CD9 is associated \\lith small CTP-bindillg proteins. Biochem Biophys Res Commun, 1991. 179(1): p. 401 -6.

23 1. Bri sson, c., et aI. , Co-localization ofCD9 and GPllh-lIla (alpha fib beta 3 il/legr;n) 01/ activated plateiel pseudopods and alpha-granule membranes. Histochem J, 1997. 29(2): p. 153-65.

232. Wessells, J., The clolling, citaracrerizatiol1 afld/Ilentiollal ww~vsis o/mllrine pregnancy specijic g~vcoproteins, in Pathology. 1999, Unifomled Services Univers ity of the Health Sciences: Bethesda, MD. p. 150.

233. Mase ll is-Smith, A. and A.R. Shaw, CD9-reglllated adhesion. Ami-CD9 mOl/oc/onal antibody induce pre-B cell adhesion to bOlle marrow fibroblasts throllgh de 1I0VO recognition ojjibl'ollectjll. J Immunol, 1994. 152(6): p. 2768-77.

234. Beauchemin. N.D., P. ; Dvcksler, G.; Go ld , P.; Gray-Owen. S.; Grunert , F.; Hammarstrom. S.; Holmes, K. Y.; Karlsson, A. ; Kuroki , M, ; Lin, S.-H.; Lucka, L.;, RedeJined nomenclature for members o/the cal'cilloemblyollie alltigell lami~v.

Exp. Cell Res. , 1999. 252 : p. 243-249. 235. Barnett . T. and W. Zi mmemlann, Workshop report: proposed nomellclafllrefor

the ca/'cil1oemh/~vonic all ligen (CEA) gene family. Tumour Bioi, 1990. 11 (1-2): p. 59-63.

109