Uterine Epithelial Cells Synthesize Granulocyte-Macrophage Colony ...

BIOLOGY

OF

46,

REPRODUCTION

Uterine

Epithelial

1069-1079

Cells

(1992)

Synthesize

Granulocyte-Macrophage

and Interleukin-6 SARAH

in Pregnant

A. ROBERTSON,2’3

GRAHAM

of Obstetrics

Departments

and

and

Nonpregnant

MAYRHOFER,4

and

and

Gynaecology,3’5

of Adelaide,

Universiy

South

Colony-Stimulating Mice1

ROBERT

F. SEAMARK5

Microbiology

Australia,

Factor

and

immunology”

5001

ABSTRACT secretion

Cytokine

by endometrial

macrophage

colony-stimulating

20-fold

250-fold

and

respectively

at estrus

harvested

mice. To determine of leukocyte and

beige

range glandular

and/or

in quantities

cells

were

similar

neutralized

of GM-CSF

supernatants identify

were

murine

IL-6

origin

of

markers

cells

the

was

with

M,

as a potent

and

through

actions

The

colony-stimulating effects

on

factors

the

of macrophages pho-hemopoietic

(CSFs)

generation,

were

chromatography 23 000-26

on target

and granulocytes cytokines are

11]. These postulated

and

pregnancy,

this

activity

is attributed

primarily

by

Granulocyte-macrophage CSF) also supernatants

appears from

lymin a

rifle decidual ventional and placental

cells [7]. GM-CSF unique sizes have

tissues

[8], and

this

factor

mRNA been

cytokine

the

stimulates

prolifer-

Januars

Received

September

This

cells

sient

South

Australia

A preliminary Maternal tions

iohn

S.A.R report

Relationship, Libbev

Colloque Eurotext,

2Correspondence: University 223

is the recipient has appeared

of Adelaide,

SA GPO

of an Australian in Molecular

INSERM

Paris

Pt’

Postgraduate

and

Cellular

during

G. Chaouat

Research

andJ.

of the Mowbray)

that

498,

Adelaide,

of Obstetrics South

Australia,

and 5001.

Award. FetoEdi-

Gvnaecologv, FAX:

that I lymorigin of

of diverse

em-

as sources

to

uterine

of

GM-CSF

cytokine regulation

in the are

from

estrous

have

measured

6 levels

61-8-

ing;

1978.

1069

and

of

with dihemo-

T lymphocytes

mated cytokine

cells

pregnancy

in the

both luminal

cytokines

a role

recently

uterine

an

luminal

cells

harvested

[15].

In the

in vivo fluid are

are

celearly

accumulation

of

following

[15].

ex-

High

from

levels

in vitro

cul-

enzymatically present

in vitro.

study,

following in

we

of uterine

GM-CSF

elevated

synthesized

mouse

rapid

in secretions

and

tranin the

during

fluid

in supernatants

bioactivity

in the

uterus

at mating

mice

analysis

with cells

in the

components

in situ

consistent

have

endometrial and

and

reacILreg-

1121. Estrogen endometrial

by endothelial

occurs

detectable

in uterine both

mRNA

reported

murine

of uterine

[13], pattern

cytokines

have

to seminal

tures

reported

that

We

endometrial Department

cells

proposed in human

midgestation

reorganization

pregnancy.

Ltd. Adelaide,

Biology

the

of B and

has been expression

of IL-6

basalis

lular

(1991).

Robertson, Box

212. (Eds.

Repromed

studies

potentially

identified

function

been

[14]. It is apparent

of GM-CSF from

culture

These have

other

been

a hybridization

expression

decidua

25, 1991. by a grant

also

and

has

revealed

posure in part

However,

have

proliferation

stromal

28, 1992.

was supported

study

[7,9].

origin

GM-CSF Accepted

postulate

revealed

and

the conceptus.

ulation of hCG synthesis regulation of IL-6 mRNA

in mu-

of conin murine

cells

fluids

factor-u.

and

epithelial

and macrophages, and participation in acute-phase tions [11]. Human placental trophoblast cells synthesize 6 and IL-6 receptors, and an IL-6-mediated autocrine

(GM-

transcripts detected

and

uterus

we

which

a

fibroblasts

from

uterine

necrosis

or

against

GM-CSF

derived

from

cells

SCID, identified

Stromal

secreted

of luminal

tumor

lL-6,

and

GM-CSF

poiesis,

has

bioactivity [6] and

GM-CSF

media

Bioassay and

uteri.

GM-CSF [101 and they may contribute production. Interleukin-6 (IL-6) is a multifunctional verse biological activities including

cells, suggestregulation of

of CSF placenta

nude,

uteri

cytokines

of conditioned

decidual

to colony-

colony-stimulating to be a component cultures of human

pregnant

and

respectively.

bryological

stimulating factor-i (CSF-1, M-CSF), which increases 1 000 fold in concentration during gestation [4]. Uterine glandular and luminal epithelial cells are the origin of CSF- 1 and synthesis is regulated by ovarian steroids [5]. The CSF-1 receptor is expressed by placental trophoblast ing a role for uterine CSF-1 in the paracrine placental development [5].

elevated>

by uterine

vitro

ation of placental cells [9]. It has been proposed phocytes responding to fetal alloantigens are

function

and other to mediate

in

midgestation of both

interleukin-3,

in both

variety of reproductive events, particularly the formation and maintenance of the placenta. Uterine tissues have been recognized for some time as potent sources of CSF [2, 3]. During

000,

cytokines

cells

from

Bioactivities

interleukin-2,

identified

proliferation,

abundance

granulocyte-

were

fluid

lymphocyte-deficient

in estrous

IL-6

harvested

mice.

INTRODUCTION their

in

The

bioassays.

intraluminal

in genetically

and

cells

mated

of the

source

specific

technique employing monoclonal antibodies uterine cell subsets in vitro. These experiments

of GM-CSF

interleukin-l,

using

of uterine

synthesized

not impaired

Epithelial

and

23000

were

isolate

source

estrous

measured

contents

a panning

to

mice.

was

cytokines

and size-exclusion

antibodies,

in pregnancy

both

used

major

from

mice (IL-6)

cytokines,

in mated

bioactivity

epithelium

and

was

as the

of cells

mated

and

interleukin-6

Synthesis

for the cytokines

uterine

and

mating,

macrophages

to those

negative

functions

important

cellular lineage

by specific and

estrous

1 of pregnancy.

epithelial

as did

IL-6,

lL-6 peaks

the other

luminal

also synthesized

from

(GM-CSF)

following

on Day

and

cells

factor

abundance

and

ILmatin

1070

ROBERTSON

vitro Day

by uterine cells 10 of pregnancy.

harvested at estrus To determine the

and on Day 1 and source of these cy-

tokines, we have examined synthesis in genetically lymphocyte-deficient mice, and also by uterine cells in vitro. The identities of the two cytokines are supported by the use of antibodies to neutralize specifically the endometrial cell-derived

bioactivities

as well

as by the

finding

that

these

cytokines have molecular weights similar to those reported for GM-CSF and IL-6 derived from other tissues. We hypothesize that uterine epithelial cells are a potent source of a variety of cytokines that include GM-CSF and IL-6, and that

this

function

ine

physiology.

is potentially

of major

importance

in uter-

ET AL.

Monoclonal antibodies (mAbs) used were TIB 107: alhyi.2 and TIB 122: aCD45 (American Type Culture Collection [ATCC], Rockville, MD), NIMP RiO: aMac-1 (A. Lopez, Hanson Center for Cancer Research, tralia), and mouse thymic stroma #12 and and MTS #24) [17, 18], which were kindly Boyd, clonal

Monash antibody

rat monoclonal kindly provided L. Sachs

AND

METHODS

Animals Adult (8-12 [CBA X C57B1]

wk) [Balb/c Fl male mice

in experiments

X C57B1] Fl female were used throughout

CBA-nu/nu, C.B-17-scid/scid

employing

mice and except

female

bg/bg

C57B1/6J-

(ARC, Perth, Australia), Eliza Hall Institute [WEHI], Melbourne, c adult (6-12 wk) females mated with

Australia), Swiss adult

(Walter and and BaIb/ male studs.

All animals

University

of Adelaide

Central caged vaginal

were

obtained

from

the

Animal House unless otherwise specified in specific pathogen-free (SPF) facilities. plug was detected was designated Day

nancy.

For

males cedure

studies

requiring

were synchronized described previously

[des-Gly’#{176},D-A1a6,-ProNEt9] Louis, MO) was injected were checked for estrous

estrous using [16]:

mice,

and were The day a I of preg-

BaIb/c

Fl

fe-

a modification of a pro20 jig of an LHRH agonist

LHRH (Sigma Chemical Co, St i.p. at 1200 h on Day -4 and mice vaginal smears on Day 0, the day

of use.

Reagents Media Island,

used NY)

included

RPMI-1640

supplemented

with

or DMEM 10%

fetal

(Gibco, calf

Grand

serum

(FCS;

Commonwealth Serum Laboratories, Melbourne, 5 X 10 M 3-mercaptoethanol, and antibiotics

Australia), (RPMI-FCS

and DMEM-FCS, respectively); RPMI supplemented FCS, 1% Nutridoma-NS (Boehringer-Mannheim,

with 0.5% Mannheim,

5 X 10’ M 13-mercaptoethanol, and and serum-free DMEM supplemented

Germany), (RPMI-NS); Nutridoma-NS,

5 x

i0

M 3-mercaptoethanol,

biotics (DMEM-NS), as specified. Recombinant Ecchericbia coli-derived the

kind

6 was murine

gift

of N. Nicola

purchased from interleukin-1

purchased MA); and

antibiotics with 1%

(WEHI).

Boehringer (IL-2) and

and

murine Recombinant

Mannheim; interleukin-3

from Genzyme Immunobiologicals recombinant human interleukin-1

rine tumor necrosis factor-a (TNFa) were British Biotechnology (Oxford, UK).

GM-CSF human

antiwas IL-

recombinant (IL-3) were (Cambridge, (IL-i) and mupurchased

from

Melbourne, GM-CSF

Aus#12 by R.

Australia. A goat poly(goat aGM-CSF) and a

antibody to murine LL-6 (rat aIL-6) [19] were by J. Schreurs (DNAX, Palo Alto, CA) and

(Weismann

Institute

of Science,

Rehevot,

Israel),

respectively.

Cytokine MATERIALS

University, to murine

Adelaide, #24 (MIS provided

Assays

GM-CSF was assayed using the GM-CSF-dependent cell line FD 5/12 [20], essentially as described by Kelso and Owens [21]. Duplicate serial 1:2 dilutions of uterine fluids or culture supernatants were incubated with 2 000 FD 5/ 12 cells in 200 p.l of RPMI-FCS. After 2 days, cultures were pulsed Heights,

with I jiCi of 3H-thymidine IL) for 6 h and harvested

using a Titretech automated cell measured in a liquid scintillation IL-6 was [22]. with

assayed

using

the

ification

of a method

Arlington fiber paper

harvester. Radioactivity beta counter.

IL-6-dependent

Duplicate serial dilutions 2 000 7TD1 cells in 200

the number of cells was version of the tetrazolium

(Amersham, onto glass

cell

was

line

7TD1

of samples were incubated jil of RPMI-NS. After 4 days,

estimated colorimetrically salt MU to formazan

described

by conin a mod-

by Mossman

[23].

Briefly,

25 p.1 of 4 mg/ml MU (Sigma) in RPMI was added to each well. After incubation for 4 h at 37#{176}C, media were aspirated carefully

and

the

formazan

product

was

solubilized

dition of 100 p.l of 95% ethanol; absorbance at 570 nm using a multiwell ELISA reader. were the

was

INFa ploying

and IL-I respectively

assayed in cytotoxicity TNFa-sensitive cell line

scribed human

by Matthews and Neale [24] and the melanoma cell line A375 as described

by admeasured assays emL929 as de-

IL-i -sensitive

by Nakei

et

al. [25]. Duplicate serial dilutions were incubated with 20 000 L929 cells in 200 p.! of RPMI-FCS and 4 p.g/ml cycloheximide overnight, or with 2000 A375 cells in 200 p.1 of RPMIFCS for 4 days. Cell lysis was measured by methyl violet uptake

(0.5%

perature). acid and

in 20%

Incorporated quantitated

methanol

for

dye was by measuring

10 mm

at room

dissolved absorbance

in

tem-

50% acetic at 570 nm

using a multiwell IL-2 and IL-3 dependent T-cell

ELISA reader. were assayed using respectively the line CTII [26] and the IL-3-dependent

IL-2cell

line

as described

Du-

32Dcl3

[27]

plicate serial dilutions CTLL cells for 3 days

by Kelso

and

Owens

[21].

of samples were incubated with 500 or 2 000 32D cells for 2 days in 200

p.l of RPMI-FCS. Cultures were pulsed and harvested as described for the GM-CSF assay. All cytokine assays were performed in flat-bottomed 96well microtiter trays (Costar, Cambridge, MA) and stan-

GM-CSF dardized against recombinant titers were determined from minimum detectable amount U/mi

defined

as the

concentration

half-maximal FD5/12 cell hands to 50 colony-forming ml stimulates marrow assay

half-maxima! [21]); for

sponse Modifiers IL-6, 5 U/mI (50 7TD1

growth);

defined

and

of GM-CSF

cvtolytic

of Uterine

and

IL-i

TNFa,

IL-6 The (50

stimulating

proliferation; equivalent units [CFU]/ml, where colony development IL-2, 1 BRMP (NIH

for

IL-6 SYNTHESIS

GM-CSF and curves. was 1 U/ml

Program) U/mi; for IL-3, U/mI defined as producing

as 50%

Collection

cytokines. dose-response for GM-CSF

AND

the

5% CO2 later and

(50

tached serted

U/mI

lected

over [28];

(25-50 via

supernatant

isolated

A blunt

then

stored

by

cutting

26-gauge

containing end of the

of 30-40

were was

were

p.l in 500 the cervical

a period they

stored

Fluid

cervix.

to a 1-mi syringe into the oviductal

ine washings were obtained age

the

near at-

RPMI-FCS was then inuterine horn and uter-

p.1 of RPMI-FCS end. Flushings sec

as by day

needle

per horn) were col-

to minimize

centrifuged

tissue

at 2 000

at -80#{176}Cprior

X

dam-

g and

to cvtokine

the

bioassay.

described

epithelial-cell-enriched tween 1000 h and

were prepared under using a modification of

by Sherman endometrial h on the

1200

[29] day

for

harvesting

an

cell population. of estrus or Day

BeI of

pregnancy, uteri from individual mice were excised, placed in cold PBS (pH 7.2), trimmed of fat, mesenterv, and blood vessels, and slit lengthwise. After another rinse in PBS, pairs of uterine horns were incubated in I ml of 0.5% trvpsin (Sigma) plus 2.5% pancreatin (bovine pancreatic, type III; Sigma)

in PBS

for

45

37#{176}C. One milliliter were agitated gently plastic

Pasteur

mm

at 4#{176}C, followed

of DMEM-FCS by passage

pipette.

The

by

45

mm

at

was added and the uteri 3-4 times up and down a

supernatant

contained

epithelial

and other liberated cells, which were pelleted at 200 x g and washed in DMEM-FCS. The cell yield was 7-14 x 10’ cells per uterus for estrous and Day I mice. Uteri excised from Day 10 pregnant mice were slit lengthwise midway between the mesometrial and the antimesometrial sides, and conceptus-derived structures were completely removed by blunt dissection by means of a dissecting

microscope.

Cells

were

prepared

by

pancreatin digestion and subsequent above. Separate populations were

processing harvested

tissue

embryonic

dissected

conceptus);

from

(2) the

(1)

between

antimesometrial

side

the

trypsmn/

described from uterine units

of implantation

x i0

cells

were

obtained

(intersites

harvested

per

nod-

cells

were

plated

in duplicate

1 ml of DMEM-FCS in 4-well multiDenmark) and incubated at 37#{176}C in

in air. Culture supernatants were centrifuged at 2 000 x g and the

at -80#{176}Cprior quantified Faulding,

to cytokine

collected supernatant

bioassay.

Adherent

after uptake of Rose Bengal Adelaide, Australia) and lysis

cells

of absorbance and peritoneal

at 540 nm. exudate

24 h was cells

dye (0.25% in 1% SDS

cells

(PEC)

in by

were

harvested from adult [Balh/c x C57BI] Fl female mice. Spleen cells were cultured at 1 X 106/ml in RPMI-FCS with 4 jig/mI phvtohemagglutinin (PHA; Sigma) for 48 h and washed; supernatants were collected after 24 h further culture at 2 X i05/ml in RPMI-FCS with ng/ml phorbol 12-myristate 13-acetate were cultured in RPMI-FCS phimurium lipopolysaccharide versitv of Adelaide, Australia). were

WEHI)

J774A.1 cells ml LPS; and

cultured (ATCC L cells

Research,

jig/mi

a procedure

4-6

X 106/ml in (Nunc, Roskilde,

measurement Spleen

cer

Cell Cultures Uterine endometrial cell monolavers sterile conditions from whole uteri

and

all tissues,

I CFU/ml; for half-maximal

activity).

horns

and

For at 1-3 dishes

were PBS;

Luminal

Individual

oviduct

tissue,

re(2-

ule.

in a bone Biological Re-

10 U/mi

1071

EPITHELIUM

(peri-conceptus); and (3) the mesometrial metrial gland gion underlying the placenta (MG). Individual nodules 3 mm in diameter) were dissected free of surrounding uterine

in our 50 CFU/

Uterine intraluminal fluid was harvested essentially described by Khurana and Wales [28]. Mice were killed cervical dislocation between 1000 h and 1200 h on the specified.

UTERINE

BY

in

4 jig/mI (PMA;

and

PHA

Sigma).

10 PEC

with 5 jig/mI Salmonella Iv(LPS; gift of I. Kotlarski, UniEL-4.BU cells (from A. Kelso, RPMI-FCS

with

10

ng/ml

PMA;

TIB 67) in Iscoves MEM-FCS with 5 p.g/ (from A. Lopez, Hanson Center for Can-

Adelaide,

LPS. In each

Australia) case,

in Ham’s

supernatants

F-12-FCS

were

with

collected

5

at 24

h from cultures with a starting density of 2 X 10 cells/mI. Supernatants were stored at -80#{176}Cprior to GM-CSF and bioassay.

IL-6

Isolation of Uterine

Cell

Subpopulations

Pancreatin/trvpsin-dissociated were cultured for 16-18

h and

uterine adherent

or

vested with 200 individual mAbs

p.M EDTA. They were then (1:2 dilution of hybridoma

Hanks’ buffered mal sheep serum

salt solution [HBSS-NSS],

tr()l) for 1 h at 4#{176}C. The panned by a modification

cell

decidual cells were

cells har-

incubated supernatant

with in

with 10% FCS and 10% or HBSS-NSS alone for suspensions were of the procedure

washed, described

Wysocki and Sato [30], using Petri dishes (Disposable ucts, Adelaide, Australia) coated with affinity-purified arat (Calbiochem, La Jolla, CA; 10 jig/mI in PBS at 4#{176}C). Monoclonal cubated in coated and the nonadherent were

detached

HBSS-NSS,

were

quantitated

above. or

GM-CSF

control

then by

Prodgoat for 24 h

antibody-labeled uterine cells were inPetri dishes for 2 h at 4#{176}C in HBSS-NSS, cells were collected. Adherent cells

by incubation in normal rat serum 20 mm at 37#{176}C) followed by forceful

Both populations the supernatants

norcon-

were were by

cultured assayed Rose

(unpanned)

cells

in

in DMEM-FCS for 24 h and for GM-CSF and IL-6. Cells

Bengal

or IL-6 synthesis

(10% pipetting.

dye

uptake

as described

by adherent, is expressed

nonadherent, in

tJ/I0s

cells.

ROBERTSON

1072 Values duction

were per

normalized control

iments,

thus

to the overall mean cytokine procells for the set of panning exper-

106

allowing

the isolated experiments, arat-fluorescein

comparison

suhpopulations mAli-labeled isothiocyanate

Australia; 1:40 in HBSS-NSS and viewed with a fluorescent proportion

of cvtokine

from cells

each were (FITC;

synthesis

experiment. incubated Silenus,

cells

in the

mixed

populations.

antibody

specifIcities

dard immunohistochemical frozen sections of Day were

incuhated

lution

of hybridoma

chambers

supernatant

mouse incubated

Denmark; visualized

(Sigma;

0.02% Tissue

5 mg/mi

cultured then for

medium harvested

was from

at 2 X 106 cells/mi 48 h in serum-free

di-

BSA and

of 0.45

ml were

on calibrated (Pharmacia) were diluted

M Tris-HCI,

Iex-GV;

Millipore),

from pooled mice. Cells

enwere

for 24 h in DMEM-FCS DMEM-NS. Eighw milliliters

and of

assayed

for

lecular weights were estimated molecular weight versus VJV,, ume), using tochrome-C Statistical Data tailed)

into

BSA,

and

BSA, ovalbumin, as standards.

GM-CSF

and

IL-6.

from a linear plot (elution volume/void a-chvmotrvpsmnogen,

Mo-

of

log vol-

and

cy-

Analysis were

analyzed

to evaluate

lummnal treatmem

by Mann-Whitney

differences

was

fluids and endometrial groups; p < 0.05 was

used

to compare

Rank

in cytokine

results are expressed as the median The Wilcoxon Signed Ranks test pairs

5 mg/mI

1-mi columns of Sephadex G-25 suby spin-gel filtration. Aliquots from 1:10 in RPMI, filtered (0.22-p.m Miland

cvtokmne

Activity

and

in

Uterine

Uterine Da

luminal 1-mated

p

greater

l,p

and

Media

fluids collected by lavage from estrous mice were assayed for cytokine bioactiv-

luminal fluid (Fig. mice, with median 9

Fluid

Luminal

Cell-Conditioned

than

TNFa activities were not IL-6 activity was detected

1) from both levels in mated at estrus

[< 2-56] U/mi 0.1), in contrast to our previous finding of elevated GM-CSF in cultures of cells from Day 1 mice [15]. Cells between

harvested from Day 10 pregnant implantation sites (inter-conceptus)

adjacent to embryonic creted both GM-CSF of cells from estrous

membranes and IL-6 at levels and Day 1 uteri

ison,

from

cells

harvested

significantly less GM-CSF similar to those in other

metrial (p < uterine

uteri, or

either from

from sites

(peri-conceptus), secomparable to those (Fig. 1). In compar-

gland

tissue

synthesized

0.05) hut quantities preparations (Fig.

of IL-6 1).

A broad range of values was obtained for both GM-CSF and IL-6 levels in mntraluminal fluids and culture supernatants. We attribute this to in vivo factors rather than to error introduced during tissue preparation or assay, CSF and IL-6 activities of supernatants derived of the two uterine horns within individual animals

since GMfrom each were not

significantly different (p > 0.1; n = 8 estrous mice). The activities of supernatants deriyed from each horn differed from the mean in an individual animal by an average of 8.2% for GM-CSF and 8.0% for IL-6. Uterine

GM-CSF

Immunodeficient GM-CSF deficient

and

IL-6

in Genetically

Mice and

mice

IL-6 were

activities

in uteri

of

compared

with

activities

mated

immunoin immuno-

GM-CSF 80

AND

IL-6 SYNTHESIS

-

A

160 60

U)

“C

a

a

In Vitro

a

a

Endometrial

a a a

-4-a

(12)

(7)

Zi

a aDAY1

EST

a

DAY1

a

C

mine the

a a

PEAlDAY

(10)

La -aa

La

10

0

a a

100

a Ui

a a

60

120

0

81

9#{128}

a

a

(10)

a

a

(7)

a

a

(7)

a

9#{128}

20

a

#{128}9 -J 40

a (12)

a

aa

-

-

a

EST

DAY1

a

a a

a

cO.1

DAY

MG

1

metrial

PERI-

GM-CSF and IL-6 bioactivity in uterine intraluminal (EST) and after mating (Day 1), and in supernatants

cell

cultures

(B and

D) prepared

from

estrous

4-

a

DAY

FIG. 1. C) at estrus

origin

of GM-CSF used

were

lEST)

and

stroma

(Fig.

3). CD45

neutrophil/macrophage

and

Mac-i

markers, cells and consistent

INTER10

Day

personal

1 uteri.

3-5% of cells and less than marker.

were 1%

Larger

40%) and 10 pregnant

labeled with were labeled

proportions

mAli with

reacted

MTS #24 (60-80%). mice were similar,

lymphocytes

or

GM-CSF and phocyte-deficient),

natural

IL-6

killer

(NK)

contribution

cells

to the

of I

>‘I>

BaIh/c

mice

mice mated that neither the

fluids strains

major

(Fig.

and of conditioned were not significantly 2)

(or

to CBA Fl T lymphocytes, sources

indeed

than

with

scid/

from immuthan those of

those

of Balb/c

GM-CSF

IL-6 in vivo, or of GM-CSF or cultures in vitro. Cells obtained

IL-6 from endometrial from hg/bg mice

sized

of GM-CSF

significantly

higher

levels

(p


‘1>

C U)

),71

(9>

>17) -

(II)

Ito

-

-‘

U

media less

intraluminal

#12

2

2

U)

I-lym-

males; Fig. 1), suggesting strongly B lymphocytes, nor NK cells are

of postmating

MIS

in the thywith these

a D

scid (T- and B-lymphocyte-deficient), and Balh/c mice were mated with Swiss outbred males. Uterine luminal flushings and uterine cell monolavers were prepared on the day of detection of a vaginal plug. The GM-CSF and IL-6 contents of lummnal nodeficient

leukocte

(>3)

postmating

responses. Female nu/nu (a/13 bg/bg (NK-cell activity-deficient),

pan

to leukocvte markers the endothelial cell

Cultures hut MTS

150

the

are

respectively.

communication).

A

to evaluate

of

immunohisto-

T lymphocytes with reactivity

12U

mice

specificities

by

of mouse uterus (I)av and NK cell marker hut h’ fibroblasts in endo-

(0

Balh/c

The

The proportion of cultured endometrial cells labeled by each mAli was determined by immunofluorescence analysis (Table I). In cultures prepared from estrous and I)av I mice,

and decidua (MG), peri-conceptus (PER)-), and inter-conceptus (INTER-) nondecidualized uteri of Day 10 pregnant mice. Symbols represent individual animals and median values are scored, with number of observations in parentheses. Symbols below the dotted line have a cytokine content below the limit of detection for the bioassay. 4Median values differ significantly from median value for estrus.

competent

.

IL-6

demonstrated

on fresh-frozen sections is a pan T lymphocyte also to be expressed

(R. Boyd,

fluid (A and of endo-

and

used to isolate populations cultures of estrous, Day I endometrial cells to deter-

and cytoplasmic suhcellular distribution in uterine glandular and luminal epithelial cells. This mAb has no reactivit- with either spleen, lymph node, or thvmic leukocytes

(7)

-a

L

EST

metrial

by Isolated Subsets of

cells in the uterus was obtained. MIS #24 reacts ithelial cells in the thvmus stroma [18] and was have specificity for an epitope with a predominantly

0

ID

-J

the

reacts with endothelial mus [17], and a pattern

-j -j

I-

IL-6 Synthesis

and

6’ells

antibodies

and

CO U,

GM-CSF

chemistry 1). Thvl.2 was found

INTER-

(12)

160

L

mice.

Monoclonal antibodies were of individual cell lineages from pregnant, and Day 10 pregnant

a (7)

4TLa

EST

C57BIk

I

20

1073

EPITIIELItJM

not IL-6, than cells from Balh/c mice (Fig. 2) and Balh/c Fl mice (Fig. I; < 0.05). Whether this was a result of the bg/bg mutation is a question awaiting studies with congenic

a

-J

a

i

(12) (12)

112)

(0

B

BY UTERINE

but

SC

C,

-.

2

a

-a-

-*--

*-

(>0> I

j2_ 8

A B

0,

U §20

D>,7>

(#{128}0) ),1)

(14)

2

-

-

))

120 (‘0)

8

2

(0)

(a)

‘5

10 9#{128} (2

--

a

C,

d

balb/c FIG. C), and

nu/nu

2. GM-CSF supernatants

scid/scid

#{176}bg/bg

-Itt-

w

-k--

balb/c

nu/nu

scid/scid

-

-

bg/bg

and IL-6 bioactivity in uterine intraluminal fluid (A and of endometrial cell cultures (B and D) from mated (Day 1) immunodeficient (flu/flu. scid/scid, and beige/beige) and immunocompetent (BaIb/c) mice. Symbols represent individual animals and median values are scored, with number of observations in parentheses. Median values differ significantly from median value for BaIb/c.

1074

ROBERTSON

TABLE

1.

GM-CSF

and

IL-6 production

by subpopulations

of endometrial

ET AL. cells

isolated

by panning

with

mAbs.

GMCSFb

mAb

% Cells’

Estrus none MIS

Adherent

lL6b % Total output0

Nonadherent

% Total Adherent

100 60-80

113

57 5

95

42

aThyl.2

20-40

2

95