Induction of cytokine expression by leukemia inhibitory factor

Induction of cytokine expression by leukemia

inhibitory factor.

P M Villiger, … , Y Geng, M Lotz

J Clin Invest.

1993;

91(4)

:1575-1581.

https://doi.org/10.1172/JCI116363

.

Biological effects of cytokines are in part determined by their interactions in the regulation of

cytokine production. This study analyzes the effects of leukemia inhibitory factor (LIF) on

cytokine expression in different cell lineages. Recombinant human LIF increases levels of

IL-1 beta, IL-6, and IL-8 mRNA in human articular chondrocytes as demonstrated by

Northern blotting. These cytokine mRNAs are detectable as early as 1.3 h after stimulation

and reach their maximum after 5 h. The LIF effects are dose dependent and of similar

magnitude to those of IL-1. By metabolic labeling and immunoprecipitation it is shown that

LIF induces synthesis and secretion of IL-6. IL-6 bioactivity in conditioned media, as

measured by the B9 hybridoma proliferation assay, is increased by LIF. Effects of LIF on

cytokine expression are not confined to connective tissue cells. By PCR it is shown that

human blood monocytes express IL-6 mRNA after stimulation with LIF. An increase in IL-6

mRNA levels is detectable 2 h after stimulation, and this starts to decline by 5 h. The

response is of shorter duration as compared with IL-1 beta. In addition to increased mRNA

expression, LIF also stimulates release of biologically active IL-6 from blood monocytes. In

synoviocytes and neuronal as well as epithelial cell lines, LIF increases IL-1 beta and IL-6

gene expression. In summary, LIF induces […]

Research Article

Find the latest version:

Induction

of

Cytokine Expression

by

Leukemia Inhibitory Factor

Peter M. Villiger, Yu Geng, and Martin Lotz

SamandRose Stein Institute

for

Research in Aging and the Department ofMedicine, University ofCalifornia, SanDiego, LaJolla, California92093

Abstract

Biological effects of cytokines are in part determined by their interactions in the regulation of cytokine production. This study analyzes

the

effects of leukemia inhibitory factor

(LIF) on

cy-tokine

expression

in

different

cell lineages. Recombinant hu-man LIFincreases levels of

IL-1i8,

IL-6, and IL-8 mRNA in human

articular

chondrocytes as demonstrated by Northern

blotting. These cytokine

mRNAsaredetectable as early as

13

h

after stimulation

and reach their maximum after 5 h. The LIF

effects

aredose

dependent

and

of similar

magnitude

tothose

of

IL-1. By metabolic labeling and immunoprecipitation it is shownthat LIF

induces synthesis

and

secretion of IL-6. IL-6

bioactivity

in

conditioned

media,

as

measured

by the

B9

hybrid-oma

proliferation

assay,

is increased

by LIF.

Effects of

LIFon

cytokine expression

are not

confined

to

connective

tissue cells. By

PCR

it is shown that human blood monocytes express IL-6 mRNA

after stimulation with

LIF. An

increase

in

IL-6

mRNA levelsis detectable 2 h

after stimulation,

and this starts to

de-cline by

5 h. The response

is of

shorter

duration

as

compared

with

IL-1iB.

Inaddition to increased mRNA expression, LIF also

stimulates

release

of

biologically active IL-6

from blood monocytes. In

synoviocytes

and

neuronal

aswellas

epithelial

cell

lines,

LIF

increases

IL-1,8

and IL-6 gene

expression.

In summary, LIF induces

cytokine expression

ina

wide

variety of

tissues.

These results suggest that

through

the induction of

cytokines,

LIFcanmodulate

inflammation,

immune responses, and

connective tissue

metabolism,

and act as a

pathoge-netic mediator

in

different disease

states.

(J. Clin.

Invest.

1993.

91:1575-1581.) Key

words:

leukemia

inhibitory factor.

cytokines

*

chondrocytes

-monocytes

Introduction

A

characteristic

property

of

cytokines is their ability

to

regulate

the

expression

ofother

cytokines,

and

this

interactioncan

medi-ate ormodulate

the

biological

effects ofa

particular cytokine.

Leukemia

inhibitory

factor

(LIF)'

isa

cytokine

thatwas

ini-tially

described

on

the basis

of its

ability

toinduce

differentia-tion

of

murine

myeloid

leukemic cells

(1).

Itinhibits

differen-Address correspondence to Dr. Martin Lotz, UCSD, La Jolla, CA 92093-0663.

Receivedforpublication 18 May 1992 and in

revisedform

18 No-vember 1992.

1.Abbreviations usedinthis paper:GAPDH, glyceraldehyde

3-phos-phatedehydrogenase; LIF, leukemiainhibitory factor; MCP-1,

mono-cytechemoattractantprotein-1; OSM,oncostatinM; SP,substanceP.

tiation of

pluripotent

embryonal stem cells (2, 3), induces the

synthesis of hepatic

acutephase

proteins

(4), promotes neuro-nal

differentiation

(5-7),

inhibits

lipoprotein lipase activity

(8),

andregulates bone

metabolism

(9-1 1 ). LIF is produced by cells of the immune system ( 12-14), thymic epithelial cells,

fibroblasts, endothelial

cells, and rat liver cells ( 15-19). We

recently

identified

humanarticular chondrocytes and synovio-cytes as

intraarticular

sources

of

LIF(20) and showed that LIF

is integrated into

the

cytokine

network through

its

induction by IL- I

and

TNF.

This

study shows that LIF shares

with

IL-1 and TNF

the

ability

to

induce cytokine production in different

cell types.

Methods

Cell isolation and culture

Chondrocytes. Cartilage was obtained at autopsy or from the

Univer-sityof California, San Diego tissue bank from donors without known history ofjointdisease. For all experiments reported here, cartilage from the femoral condyles and tibial plateaus of the knee joints was used.Cartilageslices were prepared, washed with DMEM (Whittaker MAB Bioproducts, Walkersville, MD), minced with a scalpel, and treatedwith trypsin (10% vol/vol) for 15 min in a 370C water bath. Thesamples were digested in DMEM containing 5% FBS,

penicillin-streptomycin-fungizone,and 2 mg/ml clostridial collagenase type IV

(Sigma Immunochemicals, St. Louis, MO). The cells were washed threetimes and cultured as primary chondrocytes at high cell density (4

X 106 cells per T175 flask).

Monocytes. PBMCswereisolated fromheparinized blood by Fi-coll-Hypaque density gradient sedimentation. Monocytes were iso-latedby adherence togelatin-coatedculturedishes that were

preincu-bated withautologous plasma. The plateswerewashedtocompletely

removenonadherent cells. Puck's EDTAsolution was added, the plates wereincubatedat4°Cfor 20 min, and the monocytes were collected by gentlyscrapingwitharubberpoliceman.For invitro culture, cellswere

resuspended at 106/mlin RPMI 1640containing 1% FBS,

L-gluta-mine,andpenicillin-streptomycinandplatedat

106

perwellinsix-well tissue cultureplates.

Synoviocytes. Synoviocyteswereisolated from rheumatoid arthritis

synovial tissues and culturedasdescribed (21). The cell lines were

established from the adherent primarycellpopulationsthat contain monocyte- and fibroblast-like cells.

Cell

lines

The human cell lines A549(lung carcinoma)andU373 (glioblastoma-astrocytoma)wereobtained from the AmericanTypeCulture Collec-tion(Rockville, MD)andmaintained inDMEMcontaining10% FBS.

IL-6

bioassay

The B9 hybridoma proliferationassay was usedto measureIL-6as

previouslydescribed(21).

Metabolic

labeling

and

immunoprecipitation

Chondrocytesweremaintained in T175flasksin DMEM

containing

10%FBS,L-glutamine, and antibiotics.Formetabolic

labeling,

cells wereplated in 24-well platesat adensityof40,000cells per well. Before theexperiment,cellswereserum-starved

overnight,

washed with

PBS,

J.Clin.Invest.

© The American

Society

for Clinical

Investigation,

Inc.

0021-9738/93/04/1575/07

$2.00

and incubated incysteine- and methionine-free RPMI (ICN Biomedi-cals, Inc., Costa Mesa, CA) supplemented with L-glutamine,

penicillin-streptomycin, 1%FBS, [35S]cysteine,and[35S]methionine ( 100MCi/

ml; ICN Biomedicals, Inc.) for 48 h. The supernatantswerecollected andprecleared by incubatingwithproteinASepharose (Zymed

Labo-ratories, Inc., S. SanFrancisco, CA) for 1 h. The sampleswere incu-batedovernightin the presence of 5,lof antibodytoIL-6(rabbit IgG, 1 mg/ml)on arotor, andproteinASepharose was added foran addi-tional 2 h. Theprecipitateswerewashed five times with PBS containing 0.05% Tween-80, 0.1% TritonX-100, and 1 mMPMSF followed by

twowashes in PBS. The beadswereboiled for 5 min in Laemmli sam-ple buffer and the proteins wereseparated on 12.5% polyacrylamide

gelsunderreducingconditions. Afterstaining with Coomassie blue, the

gelsweretreatedwithAmplify(Amersham Corp., Arlington Heights, IL) for 1 h,dried,andexposed to Kodak XAR film (Eastman Kodak

Co., Rochester, NY)at-70'C for 24 h.

RNA

probe preparation

The600-bp HindIII-Dral fragment of the IL-6 cDNA was inserted into thetranscription vector pGEM-3Z (Promega Corp., Madison, WI). The cDNAfor IL-8waskindly providedbyDr.K.Matsushima

(Kana-zawaUniversity, Ishikawa, Japan).A500-bpfragmentwassubcloned into the EcoRI site ofpGEM-3z,and its identitywith the template cDNAwasconfirmedby dideoxy sequencing usingT7 DNA

polymer-ase(PharmaciaLKBBiotechnologyInc.,Piscataway, NJ).For devel-opment ofaprobefor(3-actin,two 15-baseoligonucleotides (5'-CGT-CGTCGACAACGG-3' and 3'-GACCGTAGCACTACC-5') defining

a438-bp fragmentof the cDNA weredesigned. Therestriction sites EcoRI and HindIll were added at their 5' ends, respectively. After

amplification byPCR, thefragmentwasinserted into pGEM-4z. The recombinantplasmidswerelinearized and transcribed with theT7or SP6 RNA polymerase to obtain antisense probes. The probes were

labeledwith

[32P]UTP

(Amersham Corp.) and separated from

unin-corporatednucleotidesby gelfiltration(Centri-sepcolumns; Princeton

Separations, Adelphia, NJ).

Northern blot

hybridization

Primarychondrocytesweremaintained in T175 flasks for4-7d. Be-forestimulation, cellswereserum-starved for 24 h. Total RNA was extracted bythe single-step guanidinium

thiocyanate-phenol-chloro-form method(22).10-30

_,g

of totalRNA wasseparated on 1%

formal-dehyde gels,blottedontonylon filters,andcrosslinked withultraviolet

light for 5 min oneach side. The blotswereprehybridized in 50%

formamide,6x SSC, 0.5% SDS,0.1%Tween-20, and100 tgtRNA/ml for 15 minat65°C. Theprehybridizationmixture was replaced with freshsolutioncontaining106 cpm/ml ofprobe. Hybridization was per-formedovernightat65°C and followed bytwo30-min washes inlX SSC,0.1%SDSatroomtemperature and two 30-min washes in 0.1X

SSC,0.1% SDSat65°C.The damp filters were exposed to Kodak XAR filmat-70°C for 2-24 h. To confirm equal RNA load and complete transfer the 18S and 28S bandswerevisualized with ethidium bromide.

Inaddition,RNA load wasexamined by probing for

_fl-actin

mRNA.

PCR

for

cytokines

Total cellular RNAwasisolated from 106cells by the guanidinium

thiocyanatephenol chloroform method and used for reverse transcrip-tionwithM-MLVreversetranscriptase. The cDNA preparations were diluted to different concentrations for the detection of cytokine mRNAs. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH)

mRNA served asan internal control to allow comparison of RNA levels between different samples. PCR was performed on an Ericomp temperaturecycler in a 50 ,l reaction volume including 1.25 U taq

DNApolymerase, 50 pM of each primer, and - 10% of the cDNA

template.The products of 30-35 cycles ( 1 min at 94°C, 1min at 60°C,

and 2minat72°C)wereseparated by agarose gel electrophoresis and visualized by ethidium bromide staining. The PCR primers for IL-

If

weresense: GAG CTC GCC AGT GAA ATG ATG GC (position

16-38)

andantisense: GCC CTG AGT GTC GTTTTTTCG AAC

(position 402-379), defining a 387-bp fragment. The IL-6 primers were sense: CAC AGA CAG CCA CTC ACC TCT TC (position 43-65)and antisense: TGT TGA GAT GAG TAA GAC GCG TCG

(posi-tion507-484),and resulted inanamplified fragment of 465bp.The primers for GAPDH were sense: TGG TAT CGT GGA AGG ACT CAT GAC and antisense: ATG CCA GTG AGC TTC CCG TTC AGC, defining a 190-bp fragment. In some studies PCRwasperformed with[32p]dCTP in 25 cycles, which in separateexperimentswasshown to be within the linear range ofamplification for both IL-6 and GAPDH. PCRproducts were separatedon 8%acrylamide gels. The IL-6 and GAPDH bands were cut out and solubilized in scintillation liquid, and radioactivity was counted.

Reagents

The following reagents were used:

IL-lI#,

LIF, andrabbit antibodyto

human IL-6 (R&D Systems, Inc., Minneapolis, MN). Endotoxin con-tent of the cytokine preparations was<0.1 ng per 1 ,Agof IL-I or LIF protein. LPS from Salmonella Minnesota and PMA (Sigma Immuno-chemicals).

Results

LIF

induces IL-6 and IL-8

mRNA

expression

in

chondrocytes.

Primary human articular chondrocyteswereserum-starvedfor 24 h to avoidinterference ofserum componentswithmRNA induction by the specific cytokines. Cells were then stimulated with LIF and RNA was analyzed by Northern blotting. Fig.1A showsthat IL-6 and IL-8 mRNAare notdetected in unstimu-lated chondrocytes. Stimulation with LIF for 4 h causeda dose-dependent induction of both cytokines. The IL-6 and IL-8 mRNAshad the expected sizes of 1.3 and 1.7 kb, respectively. LIFeffects on IL-6 and IL-8 mRNAs are already detectable 1.3 hafter stimulation (Fig.1B) andlevels increase further for 5 h.

A Co 50 20

10

IL-6 Ig p i

IL-8

P-actin

B

LIF

1.3 2.5 3.8 5 2.5Co 2.5IL-1 Co5 IL-15

IL-6

IL-8 m

P-actin *H

Figure1. LIFinduces IL-6and IL-8 mRNAexpressionin chondro-cytes.Primaryhumanarticularchondrocyteswereserum-starved and leftinmedia alone(Co, control)orstimulated with LIForIL-flfor

4 h(A).B showsRNAfromcells thatwerestimulatedwith IL-

1O

or

LIFfor the timeperiodsindicatedinhoursabove thelanes.Total cellularRNAwasanalyzedfor IL-6andIL-8expression byNorthern

blotting.Thesamefiltersweresubsequently probedfor

fl-actin

to

1 23 45 6 789

ILiB

GPDH

-IL-1 is a well-characterized and potent inducer of IL-6 and IL-8 in chondrocytes (23). Fig. 1Bshows the results of a com-parative analysis of LIF and IL-1,8. After 2.5 or 5 h stimulation, LIF

induced

quantitatively increases in IL-6 and IL-8 mRNA levels

similar

to IL- I.

Although IL-I

is

an

important

regulator of chondrocyte

function,

it is not clear whether chondrocytes express the IL-I

gene. We thusanalyzed IL-

_1#

mRNA

expression

in chondro-cytes.

Fig.

2 shows that IL-18 mRNA

is readily

detectablein activated primary human articular chondrocytes. Only a faint signal isseen in unstimulated cells, but this is markedly

in-creased

after

chondrocyte

activation with

cytokines. IL-13 mRNA levels were

induced

by IL-

_1#,

andthis is consistent

with

IL-1

autoinduction,

which has been demonstrated in other cell types. Chondrocyte

stimulation

with LIF caused a

similar increase

in IL- 13 gene

expression. Fig.

2 also shows that

IL-6

can

induce

IL- 13 mRNA in chondrocytes, while the growth

factors TGFB,

PDGF, and bFGF have no detectable

effect.

These results show that LIF induces 6, 8, and

IL-1B

mRNAin

primary

chondrocytes and that the magnitude

and

kinetics of its effects

are

similar

tothose

of

IL-1B. LIF

effects

on

IL-6

expression in human

blood

monocytes.

LIF regulates

differentiation of myeloid

leukemia cells and mononuclear phagocytes express LIF receptors (24). As mono-cytesare an

important

source

of cytokines,

we

tested whether

LIF can

induce the

expression ofproinflammatory cytokines

in

this

cell type.

Fig. 3 shows

the

results from PCR

analysis for

IL-6

expression.

Low levels

of IL-6

mRNA weredetected in

1 23 4 5 67 89

Figure 2. Regulation of IL-

I/

expression in chon-drocytes. Primary articular chondrocytes were stimulated in serum-free media for 4 h with cyto-kines and growth factors. RNA was isolated and cDNA was prepared by reverse transcription and amplified by PCR in the presence of IL-I

_#-specific

primers. Aliquots of the same chondrocyte cDNA were amplified in the presence of primers for GAPDH (to serve asacontrolfor the amount of template in thedifferent samples). PCR products were separated onagarose gels, stained with ethid-ium bromide, and visualized under ultraviolet light. Lane 1, 10 ng/ml PDGF; lane 2, 10 ng/ml bFGF; lane 3, 10 ng/mlIL-1,8;lane 4, 10 ng/ml IL-6; lane 5, 10 ng/mlLIF; lane 6, 10ng/ml

TGF-#

1;lane 7, 500 U/ml TNFa; lane 8, 5%FBS; lane 9, serum-free media control.

adherent monocytes. A strong increasewas seenwith

LPS,

a

known

inducer of

IL-6. LIF dose

dependently

stimulated IL-6

mRNA

levels.

LIFeffects were

already

apparent

by

2 h and

started

to decline

by

5 h. The increase in

steady-state

IL-6

mRNA

levels by

LIF was greater than

that

induced

by

IL-1,B

in thesamedose range.

PCR

analysis with 32P-labeled

nucleotides showed that IL-6 mRNA

increase

-

10-fold

in the presence of LIFat50 and10

ng/ml under conditions where

the levels of GAPDH

radiola-beled PCR

product

were constant

(not shown). Sequencing

of

PCR

product that

was obtained with the IL-6

primers

con-firmed that

it had the

expected nucleotide

composition

ofthe

IL-6

gene between

positions 43 and 507.

IL-6 is

de novo

synthesized

in

response

to LIF.

Having

shown that LIF can induce

cytokine

mRNA

expression,

we

analyzed

whether gene

expression

resulted ina

corresponding

increase

in

protein synthesis. Subcultured (passage 1)

or

pri-mary

chondrocytes

were

serum-starved for

24 h

and then

stimu-lated

with

LIFat

concentrations between 10

and 50

ng/ml.

Conditioned media

were harvested after 48 h and IL-6 was

immunoprecipitated and analyzed by

SDS-PAGE. The effects

of IL-I/

are shown for

comparison.

Inboth cell

populations

(subcultured cells,

lanes

1-5;

primary chondrocytes,

lanes

6-10)

LIF

induced dose-dependent synthesis

and

secretion

of

IL-6

proteins which

weredetected as the

characteristic

bands between

20

and

30

kD,

representing

the

different molecular

forms of IL-6

(Fig.

4).

The

relative

intensity of

these

bands is

similar

in IL-

1I-

orLIF-treated cells.

Thus,

LIFnot

only

leads

Figure 3.LIFinduces IL-6mRNA

expression

in monocytes.Monocyteswerestimulated with

LIF,IL-I,orLPS for 2or4h.IL-6 mRNA

wasmeasuredbyPCRasdescribed in

Fig.

2. Lanes1 and 7,

unstimulated;

lanes2 and8,

1 ,ug/mlLPS;lanes 3 and9, 10

ng/ml

IL-1;

lanes 4 and10, 50ng/ml LIF;lanes5and11,

10ng/ml LIF;lanes 6and12, 1

_{ng/ml LIF;}

lanes 1-6, stimulation for 2h;lanes 7-12,

stimulation for4h.Upper

panel

shows the IL-6PCRproduct.Lower

panel

shows the result from

amplification

of

aliquots

of thesame

LIF

'

0

20 10

IL-1 Co

LIF

50

20

10

I

.. _L

_bs

_*._

_r..

...

_{>i_.''}

.&.2....

_

_{''_F}

_

__ si. _ _ __. _

_

_.s

.

_

24.0

kD-20.1

kD-Figure 4. IL-6synthesisandrelease by LIF-treatedchondrocytes. Human articular chon-drocytesasprimary(left)orsubcultured

(right) cellswerestimulatedwithLIF

(con-centrations innanogramspermilliliter)or

IL-1B (1 ng/ml) and metabolically labeled for 48 h. Cellscultured in media alonewereincluded ascontrol(Co).IL-6wasimmunoprecipitated

fromconditioned media and separatedon

SDS-PAGEunder reducing conditions. Gels

weredried andnewly synthesizedIL-6proteins werevisualized by autoradiography.

toarapid expressionof IL-6 mRNA but it also induces

synthe-sis and secretion of IL-6 proteins.

LIF increases IL-6

bioactivity

in

conditioned media.

To

testwhether IL-6 in conditioned media from LIF-treated cells

wasbiologically active, IL-6activity wasmeasured in the B9

hybridoma proliferationassay. LIFinducedadose-related

in-creaseinIL-6bioactivityin 24-hchondrocytecultures tolevels

similartothosefound inIL-1I3-stimulatedcultures(Fig.5).At

lower doses (0.1 ng/ml)

IL-Ifl

appeared to be more potent

thanLIF.

Low levels of IL-6 activitywerefound in 24-h conditioned

media from adherent blood monocyte cultures.LIFcauseda

maximal 10.5-fold increase in IL-6 levels. The maximalIL-1

#

effectwas a 13.4-fold stimulation of IL-6. Collectively,these

results showthat LIF inducessynthesisand secretion of biologi-callyactive IL-6 inmonocytes andchondrocytes.

Since LIFcanstimulate IL-l expressionandIL-1 isa

po-tent inducer ofIL-6, it was _possible that the LIF effect was

dependent on IL-1. Fig. 6 shows that the LIFinductionwas

almostcompletely inhibited by antibody to LIF but not by

anti-IL-I orpreimmunerabbitIgG.TheantibodytoIL-1 was

biologicallyactive since itwasabletoreduce theIL-1effecton

IL-6 expression. Thesefindings suggest that the induction of

IL-6 isadirect function ofLIF. Inaddition,theseresults also

demonstrate that the induction of IL-6wasnotduetothe pres-enceof endotoxin in the LIFpreparationused.

LIF

induction

ofcytokines

in

other cell

types.

To

determine

whether the effects ofLIFoncytokineexpressionarerestricted

toparticularcelltypesorrepresentageneralresponsein

differ-ent tissues, human rheumatoid synoviocytes, glioblastoma,

andlung epithelial cellswerestimulated withLIFand probed

for theexpressionofdifferent cytokine mRNAs.LIFinduceda

dose-related increase in IL-6 mRNA expression in

synovio-cytes, in theepithelial cell line A549, and in the glioblastoma

celllineU373 (Fig. 7). The increase in mRNA levels inall cell lines inducedbyLIFand IL-1If wasof similar magnitude.LPS hadonlyweakornoeffectsonIL-6 expression in synoviocytes

andA549andU373 cells,whichsuggeststhattheabilityof LIF

toinducecytokine mRNAsisnotduetothepresenceof small

amounts ofendotoxin. Inaddition toIL-6,LIF also induced

the expression ofmRNAs for IL-1,B andthe monocyte

che-moattractant protein- in these different cell types (not shown).

For betterquantificationof the LIF effectsonIL-6mRNA levels, PCR for IL-6 and GAPDHwasperformed in the

pres-enceof [32p]dCTP for 25 cycles, which inseparateexperiments

wasshownto be within the linearrange ofamplification for

both IL-6 and GAPDH. The results inFig. 8,whichwere

nor-malizedonthe basis of the counts forGAPDH,show thatLIF

caused adose-dependentincrease in IL-6 mRNA whichwas

10.3-, 11.2-, and 2.7-fold, respectively, atthe three different concentrations of LIF used.

Discussion

This study showsthat LIF inducesexpressionof

proinflamma-torycytokines. Human articular chondrocytesexpressIL-

1#,

IL-6, andIL-8 mRNA andsynthesizeandsecretebiologically

active IL-6 in responseto LIF. Similar effectswerefoundin

human blood monocytes,synoviocytes,andepithelialand

neu-ronal cell lines. Through the induction ofthesecytokines in diverse cell types LIF has thepotentialtoinitiate and propagate

inflammatoryandimmunologicalresponses.

LIF isa cytokine thatregulatesdifferentiation ofa wide variety ofcelltypesincludingembryonicstemcells and hemato-poieticand neuronal cells (25). More recently,itwas

recog-nized thatLIFalsoregulates bone metabolism in vitro and in vivo (9-11 ) and that it induces expression of hepatic acute

phase proteins(4). Other functions include the inhibition of lipoprotein lipase activity (8) and the increase in neuronal

sub-stanceP(SP) expression (26).

AlthoughLIFdoesnotsharesequencehomology with IL-6,

these twocytokines have somesimilar functional properties.

Bothaugment proliferation ofhematopoietic stemcells (27)

andinducedifferentiation of murine Ml leukemic cells (28). Furthermore,IL-6 andLIFinduce expression ofasimilarsetof

acute phase proteinsin hepatocytes (4, 29). These findings

could beexplained byareciprocal induction of thetwo

cyto-kines, but in hematopoietic cells this mechanism has been

ruledoutusing neutralizing antibodies (27).Another explana-tion forfunctional similarities is the possibility that LIFand IL-6stimulatesimilar intracellular signals. This hypothesiswas

supported by the characterization ofthe LIF receptor(30),

which hasstructural homologywith the IL-6 signal transducer,

gpl30.Furthermore, thesameimmediate earlypatternofgene

expressionhasbeenobserveduponinduction of myeloid

E

'a

2000

-

1000-0

3000

-iZ3~

CL

CHONDROCYTES

anti-IL-1

- IL-A

preimm. IgG

anti-LIF

anti-IL-1

LIF

Contr

CTR 1

j0

1 0.1 1 0 1 0.1

LIF ll

IL-18B

Figure5.Chondrocytes andmonocytes secretebiologically active IL-6 inresponsetoLIF.Conditioned mediawerecollected from 24-h

cultures ofcytokine-stimulated chondrocytes andmonocytes.

Con-centrations ofIL-1,BandLIFareshown innanogramspermilliliter. The samplesweretestedinthe B9 hybridoma proliferationassayfor

IL-6activity.IL-6wasquantifiedonthe basis ofastandardcurve

us-ing recombinant human IL-6. Resultsareexpressedas meanvalues

fromonerepresentative experiment (total number of experiments

=4) performedintriplicate.

kemia celldifferentiationby IL-6 andLIF(31 ). Oncostatin M (OSM) (32) has effects similarto those of IL-6 and LIFon

hepatocytesandmurinemyeloid leukemia cells and, although

this hasnotbeenformally demonstrated,it isconceivable that

other cell types show similarresponsestoLIFand OSM. This notion isbasedontherecentdemonstrationthat the high affin-ityreceptorforLIF, which is composed of the IL-6 signal

trans-ducergp130andtheLIF receptor-achain,is alsoahigh affinity

OSM receptor(33).

IL-I and IL-6 share somebiological effects, but thereare

alsoimportantfunctional differences. These include the

induc-0 400 800 1200 1 600

IL-6(pg/ml)

Figure 6.LIF induction ofIL-6isnotIL-I dependent. 5 ng/mlLIFor

1 ng/mlIL-Iwaspreincubated with specific neutralizing antibody

orpreimmunerabbitIgG (allat 10Ag/ml)orinmediaonly.These mixtures ofcytokinesandantibodiesand LIForIL-I that hadnot beenpretreated with antibodywerethen addedtochondrocyte

cul-tures.IL-6activityin 24-hconditioned mediawasdeterminedinthe

B9assay. Resultsrepresentmean±SEM oftwoexperiments

per-formedintriplicate.

tion ofproteases by IL-1 (34) but theinduction ofprotease

inhibitorsby IL-6 (35)or amarked differenceinthe induction

of otherproinflammatory cytokines.IL-I butnotIL-6induces

IL-8 andmonocytechemoattractantprotein-I (MCP-1) (23,

36). AlthoughLIFshares functionalpropertieswithIL-6,

sev-eralfindingsindicate that thequalityof the LIF effects ismore

proinflammatoryand similarto IL-1.LIFinhibitslipoprotein lipase activity (8), whichcanleadtocachexia(10).IL-I and

LIF, butnotIL-6, increase neuronal SP expression (26). We recentlyshowed thatLIFinducesexpressionofcollagenaseand

stromelysin byhuman articularchondrocytes,and that it does

notstimulate theexpressionof tissue inhibitor of

metallopro-teinases(Lotz, M., P. M. Villiger, R. Melton, and V. Ganu, manuscriptinpreparation). Furthermore, LIF, but notIL-6, induces expressionof MCP-1inchondrocytes (36).Thisstudy

providesevidence that theproinflammatoryactions of LIF in-clude induction of the cytokinesIL-1 and IL-6aswellasthe

neutrophil chemoattractant protein IL-8. These findings stronglysupport the notion that LIFcandirectlyandindirectly

initiate inflammatory events and suggest a synergy between LIF and IL-I since bothcytokinescaninduce eachother's

ex-pression.

Arthritis isoneexample ofmanydiseases where this LIF

effectcancontributetopathogenesis. Cartilage is the final

tar-get ofjoint inflammation. Independent from the underlying

diseaseprocesses, inflammationresultsindestruction of carti-lage leading to impairment ofjoint function and arthrosis. Chondrocytesare nowknowntoactively participatein

regula-tion ofinflammatory reactions through the production ofa

broadrangeofcytokinesandgrowthfactors(20, 23, 36, 37).It

wasnotclear whetherchondrocytes also produceIL-1,which is

acritical catabolic factor forcartilagesince it inhibits the

syn-thesis ofglucosaminoglycansand stimulates theproductionof

metalloproteinases. We show here that chondrocytes indeed

expressthe IL-1 geneand that IL-Icanautoinduce itsmRNA

inthis cell type. Moreimportantly,LIFinducesIL-13 in

chon-I

preimm.IgG

A

i

I r

Synov.

Chond.

123456

123456

123456

123456

IL-6

GPDH

Figure7.Effects ofLIFon

cytokine

mRNAexpression indifferent celltypes.A549cells, U373cells,

chondrocytes,

and humanrheumatoid synoviocyteswereplatedinsix-well plates, grown to confluency, and serum-starved for 48 h. Thecellswerethenstimulatedwith LIForIL-LSI

for4h.RNA wasextracted andanalyzedfor IL-6 mRNA levels by PCR. Lane 1, control; lane 2, 50 ng/ml LIF; lane3, 10ng/ml LIF;lane4,

1 ng/ml LIF; lane 5, 10 ng/ml IL-Id; lane 6, 100ng/mlLPS.

drocytes.

Thus,

LIF can

directly stimulate cartilage destruction

via the induction of proteinases and indirectly through the in-duction of IL- 1.

Cartilage

thus isa sourceof two

cytokines

that induce connective tissue catabolism. The

function

of thesetwo

cytokines

is

potentiated since they induce each other's

aswell

as

their

own

production (20).

This

complex

interaction of the two

cytokines

raises the

possibility

that the

effect of

one may be

mediated

via the induction of the other factor.

However,

the

rapid induction of IL-6 and IL-8 by

IL- 1

and

LIF

which

occurs

within

2

h, and the fact that

antibody

to IL-l

did

notinhibit

LIF-induced IL-6

production,

suggest

that these

are

probably

direct

LIF

effects. This

was

supported by

experiments

where

neutralizing antibody

toIL-

I1 did

notinhibit the

induction

of

IL-6 by

LIF.

LIF

is also expressed by

synovial

tissue cells.

We

recently

found

expression

of

high

levels

of

LIF mRNA

in

responseto

stimulation with

IL-1 in

synovial

organ

culture

(20)

where mRNA was

primarily

localized

in

interstitial cells.

For

the

anal-ysis of

LIF

effects

on

cytokine

expression

in

synoviocytes

we

did

not

detect

constitutive levels of

cytokine mRNAs,

which can

be

found in

freshly

isolated

rheumatoid

synovium.

Inthis

study

we

used cells that had been cultured in

the

absence

of

serum

for 24-48

hto

avoid contribution of

serum

factors

to

the

LIF

effect

on

the

induction of other cytokines.

As LIF can

induce expression of the

two

chemoattractant

proteins,

IL-8

and

MCP-

1, an

indirect role of

LIFin

recruitment

ofinflamma-tory

cells which

are

typically

present in

joint inflammation

can

be

proposed.

Furthermore,

LIF

could

enhance the

proinflam-matory

functions of

the

invading

monocytes as

it

induces

cyto-kine

expression in

these cells.

With its effects

onchondrocytes and

synoviocytes

LIF can

contribute

to

pathogenesis of arthritis. However, through

the

ability

to

stimulate cytokine production in mononuclear

phago-cytes, LIF

is

a

pathogenetic

mediator in diverse

disease states. LIFhad

originally

been

described

as aninducer of monocyte

differentiation of

the

murine

Ml

myeloid

leukemiacell line

(

1

). However, although

LIFreceptors have been found on

mac-rophages,

monocytes, and

their

precursors (24), no effect of LIF on mature monocytes

has

been described. The present

findings of

LIFinduction of IL-6 mRNA and IL-6 activity are

the

first demonstration of

LIF

effects

onmature monocytes.

Based on these findings we have more recently shown that LIF canalso stimulate HIV replication in mononuclear phagocytes (Broor,

S.,

A. B. Kusari, B. Zhang, P. Seth, D. D.

Richman,

D. A.

Carson,

W.

Wachsman,

and M.

Lotz,

manuscript

in prep-aration). The ability of LIF to stimulate cytokine

expression

appears not tobe restricted to a particular cell type. In addition to

chondrocytes

and monocytes, we observed similar effects in

synoviocytes

and

epithelial

and neuronal cell

lines,

which sug-gests that LIF is a ubiquitous inducer of cytokines.

Insummary, LIF is a potent inducer of

cytokines

such as IL-13,

IL-6, MCP-1,

and IL-8 in a broad spectrum ofcell types. Through these effects that have thus far only been identified for IL-1 and

TNF,

LIF can

regulate

inflammatory

and immune responses and promote

connective tissue

catabolism.

LIFthus

2000-0

:::'::: .:::: . ':::''''A'

'-E_C. 1000

,,,,Eo

-...

_{I....|...:.0-.-. * S'''*a'*:.}

con000t-r_{...t...r--....}...L5L 1I '.-.,''.'.'.".,';.'... ,"'"''""""" :" ...

...

~~~~~~~~~~~~...--.-....

~~~~~~~~~...

j

'~~~~~~~~~~~~.:.:.:

Contr LIF 50 LIF 10 LIF 1

Figure 8.Quantification of LIF effects on IL-6 mRNA expression.

Chondrocyteswerestimulated for4hwith LIF in the concentrations indicated (nanograms per milliliter). RNA was isolated and PCR

wasperformed for IL-6 and GAPDH in the presence of

[32P]dCTP.

The PCR products were separated on acrylamide gels and excised, andradioactivitywasquantified by scintillation counting. Results are shownascountsperminute of IL-6 PCR product, which were nor-malizedonthebasis of the counts for GAPDH.

1580 P.M. Villiger, Y.Geng,and M. Lotz

has the potential

to

be

a

pathogenetic

mediator in

diverse

dis-ease states.

Acknowledgments

We thankJacquelineQuach and Jinae Shin for expert technical assis-tance. Drs. KurtBenirschke, Paul Wolff,and Glenn Greenleaf pro-vided thecartilage samples that were used in this study.

This work was supported by National Institutes of Health grants AR-39799,AG-07996, and RR-00833.

References

1.Hilton, D. J., N. A. Nicola, N. M. Gough, and D. Metcalf. 1988. Resolution andpurificationof three distinctfactorsproduced by Krebs ascites cells which havedifferentiation-inducing activity on murine myeloid leukemic cell lines. J. Biol. Chem. 263:9238-9243.

2.Williams, R. L., D. J. Hilton, S. Pease, T. A. Willson, C. L. Stewart, D. P. Gearing,E. F.Wagner, D.Metcalf,N.A.Nicola,and N. M.Gough. 1988. My-eloid leukaemiainhibitory factor maintains the developmental potential of em-bryonicstemcells. Nature(Lond.).336:684-687.

3.Smith, A. G., J. K. Heath, D. D. Donaldson, G. G. Wong, J. Moreau, M. Stahl,andD. Rogers. 1988. Inhibition ofpluripotential embryonicstem cell differentiation bypurified polypeptides.Nature(Lond.).336:688-690.

4.Baumann, H., and G. G. Wong. 1989. Hepatocyte-stimulatingfactorIII shares structural and functionalidentitywithleukemia-inhibitory factor.J. Im-munol. 143:1163-1 167.

5.Yamamori,T., K.Fukada, R. Aebersold, S. Korsching, M. J. Fann, and P. H. Patterson.1989. Thecholinergicneuronaldifferentiation factor from heart cellsis identical toleukemiainhibitory factor. Science(Wash.DC).

246:1412-1416.

6.Yamamori,T.1991.Localization ofcholinergic differentiation factor/leu-kemiainhibitory factor mRNA in the rat brain and peripheral tissues. Proc.Natl. Acad.Sci. USA88:7298-7302.

7.Murphy, M., K.Reid, D. J.Hilton,and P. F. Bartlett. 1991.Generation of sensoryneuronsis stimulated by leukemiainhibitory factor. Proc.Natl. Acad. Sci. USA. 88:3498-3501.

8.Mori,M., K.Yamaguchi,and K. Abe. 1989. Purificationofalipoprotein lipase-inhibitingprotein produced by a melanoma cell line associated with cancer cachexia.Biochem.Biophys.Res.Commun. 160:1085-1092.

9. Abe, E., H. Tanaka,Y.Ishimi, C. Miyaura, T. Hayashi, H. Nagasawa, M. Tomida,Y.Yamaguchi, M. Hozumi, and T. Suda. 1986. Differentiation-induc-ing factorpurifiedfrom conditioned mediumofmitogen-treated spleen cell

cul-turesstimulates boneresorption.Proc.Natl.Acad.Sci. USA83:5958-5962. 10.Metcalf, D., and D.P.Gearing. 1989. Fatal syndrome in mice engrafted with cells producing high levels of the leukemia inhibitory factor.Proc. Nat!.

Acad. Sci. USA 86:5948-5952.

11.Reid,L.R., C. Lowe, J. Cornish, S. J. Skinner,D.J.Hilton,T. A.Willson, D. P.Gearing, and T. J. Martin. 1990. Leukemiainhibitoryfactor:anovel bone-activecytokine.Endocrinology. 126:1416-1420.

12. Godard, A., H.Gascan, J. Naulet, M.-A. Peyrat, Y. Jacques, J.-P. Soulil-lou, and J. F. Moreau. 1988. Biochemicalcharacterizationandpurification of HILDA,ahumanlymphokine active oneosinophilsand bonemarrowcells. Blood.71:1618-1623.

13. Anegon, I., J.F.Moreau,A.Godard,Y.Jacques, M.A.Peyrat, M. M. Hallet, G. Wong, and J. P.Soulillou.1990.Productionof human interleukin for

DAcells(HILDA)/leukemia inhibitoryfactor(LIF) byactivated monocytes: productionofhumaninterleukin forDAcells(HILDA)/leukemia inhibitory factor (LIF) by activated monocytes.Cell.Immunol. 130:50-65.

14.Le,P.T., S.Lazorick,L. P.Whichard,Y.C.Yang,S.C.Clark,B.F.

Haynes, and K. H.Singer. 1990. HumanthymicepithelialcellsproduceIL-6,

granulocyte- monocyte-CSF, and leukemia inhibitory factor. J. Immunol. 145:33 10-3315.

15.Tomida,M.,Y.Yamamoto-Yamaguchi,and M.Hozumi.1984. Purifica-tionofafactorinducingdifferentiation ofmousemyeloidleukaemicMlcells from conditioned medium of mouse fibroblast L929 cells. J. Biol. Chem. 259:10978-10982.

16.Koopman,P.,and R. G. H. Cotton. 1984.Afactorproducedbyfeeder cells which inhibitsembryonalcarcinoma cell differentiation: characterization andpartial purification. Exp.Cell Res. 154:233-242.

17.Lubbert, M.,L.Mantovani,A. Lindemann,R. Mertelsmann,and F. Herrmann. 1991.Expressionof leukemiainhibitoryfactor isregulatedin human mesenchymalcells.Leukemia(Baltimore).5:361-365.

18.Smith,A.G.,and M. L.Hooper. 1987. Buffaloratliver cellsproducea

diffusibleactivitywhichinhibits thedifferentiationof murineembryonal

carci-nomaandembryonicstemcells. Dev. Biol. 12:1-9.

19.Gascan, H.,I. Anegon,V.Praloran,J.Naulet,A.Godard,J.P.Soulillou,

andY.Jacques. 1990.Constitutiveproductionof human interleukin for DA

cells/leukemia inhibitoryfactorbyhumantumorcelllines derived from various tissues. J.Immunol.144:2592-2598.

20.Lotz,M.,T.Moats,and P.M.Villiger.1992. Leukemiainhibitoryfactor is expressedincartilageandsynovium.Potential contributionstothepathogenesis

of arthritis. J.Clin.Invest. 90:888-896.

21.Guerne, P.-A.,B. L.Zuraw,J. H.Vaughan,D.A.Carson,and M.Lotz. 1989.Synoviumas a sourceof interleukin-6 in vitro. Contributiontolocal and systemic manifestationsof arthritis. J.Clin.Invest.83:585-592.

22.Chomczynski, P.,andN.Sacchi. 1987.Single-stepmethodofRNA isola-tion by acidguanidiniumthiocyanate-phenol-chloroformextraction. Anal. Bio-chem. 162:156-159.

23.Lotz, M.,R.Terkeltaub,and P. M.Villiger.1992.Chondrocytesandjoint inflammation:expression ofIL-8 in responsetopeptide regulatoryfactors and proinflammatoryagents. J.Immunol. 148:466-496.

24. Hilton,D.J., N.A.Nicola,and D.Metcalf. 1988. Specific bindingof murineleukemiainhibitory factortonormal andleukemicmonocyticcells. Proc. Natl.Acad.Sci. USA 85:5971-5975.

25.Gough, N. M., and R. L.Williams. 1989. Thepleiotropic actions of leukemiainhibitoryfactor. CancerCells(Cold Spring Harbor). 1:77-80.

26.Freidin,M., and J.A.Kessler. 1991.Cytokine regulationof substanceP

expressioninsympatheticneurons.Proc.Natl. Acad.Sci. USA.88:3200-3203. 27.Leary,A.G., G. G. Wong, S. C. Clark,A.G.Smith,andM.Ogawa.1990. Leukemiainhibitoryfactordifferentiation-inhibiting activity/humaninterleukin forDAcells augmentsproliferationof humanhematopoieticstemcells.Blood. 75:1960-1964.

28.Metcalf,D. 1989.Actions and interactions ofG-CSF,LIF, and IL-6on

normal and leukemic murine cells.Leukemia(Baltimore).3:349-355. 29.Kordula, T.,H.Rokita,A.Koj,W.Fiers,J.Gauldie,and H. Baumann. 1991.Effects of interleukin-6 and leukemiainhibitory factorontheacutephase response and DNA synthesisin cultured rat hepatocytes. LymphokineRes. 10:23-26.

30.Gearing,D.P.,C. J.Thut,T.VandeBos,S. D.Gimpel,P. B.Delaney,J. King,V.Price,D.Cosman,and M. P.Beckmann. 1991.Leukemia inhibitory factorreceptor isstructurallyrelatedtotheIL-6signaltransducer,gpl30.EMBO (Eur. Mol. Biol.Organ)J. 10:2839-2848.

31. Lord, K. A.,A.Abdollahi,S. M. Thomas,M.DeMarco, J. S. Brugge,B.

Hoffman-Liebermann,and D.A.Liebermann. 1991. Leukemiainhibitoryfactor andinterleukin-6triggerthesameimmediate earlyresponse,including tyrosine phosphorylation,uponinduction of myeloidleukemia differentiation. Mol.Cell. Biol. 11:4371-4379.

32. Rose,T. M.,and A.G. Bruce. 1991. Oncostatin Misamemberofa

cytokine familythat includesleukemia-inhibitory factor, granulocyte

colony-stimulatingfactor,andinterleukin-6. Proc.Nat!.Acad.Sci. USA. 88:8641-8645. 33.Gearing,D.P.,M. R.Comeau,D. J.Friend,S. D.Gimpel,C.J.Thut,J. McGourty,K. K.Brasher,J.A.King,S.Gillis,B.Mosley, S. F.Ziegler,andD.

Cosman.1992. The IL-6signaltransducer,gpl 30: and oncostatin M receptor and affinityconverter for the LIF receptor. Science(Wash. DC). 255:1434-1437.

34. Mizel,S. B., J. M. Dayer, S.M.Krane, and S.E.Mergenhagen. 1981. Stimulationof rheumatoidsynovialcellcollagenaseandprostaglandin produc-tionbypartiallypurified lymphocyte activatingfactor(interleukin-1).Proc.Natl. Acad.Sci. USA.78:2474-2477.

35.Lotz, M.,and P.A.Guerne. 1991.Interleukin-6 induces thesynthesis of tissueinhibitor ofmetalloproteinases-l /erythroid potentiating activity

(TIMP-1/EPA).J. Biol.Chem. 266:2017-2020.

36.Villiger,P.,R.Terkeltaub,and M.Lotz.1992.Monocytechemoattractant protein-1(MCP-I)expressioninhumanarticularcartilage.Inductionbypeptide regulatoryfactorsanddifferential effects ofdexamethasone and retinoicacid.J.

Clin.Invest.90:488-496.

37.Guerne,P.A., D. A.Carson,and M. Lotz. 1990. IL-6production by

human articularchondrocytes:modulation ofitssynthesis by cytokines, growth