Role of lymphotoxin in expression of interleukin 6 in human fibroblasts Stimulation and regulation

(1)

Role of lymphotoxin in expression of interleukin

6 in human fibroblasts. Stimulation and

regulation.

M Akashi, … , M Saito, H P Koeffler

J Clin Invest.

1990;

85(1)

:121-129.

https://doi.org/10.1172/JCI114401

.

IL-6 is a cytokine with a number of biological functions, including stimulation of

immunoglobulin synthesis and proliferation of early hematopoietic stem cells. We showed

that lymphotoxin stimulated accumulation of IL-6 mRNA in human fibroblasts (W138) in a

dose-responsive fashion; tumor necrosis factor-alpha (TNF-alpha) was about threefold more

potent than lymphotoxin. Further experiments suggested that stimulation by lymphotoxin

was independent of protein kinase C activity, did not require new protein synthesis, and was

at least in part a result of increased stabilization of IL-6 mRNA. t1/2 of the IL-6 transcripts

increased from 0.3 h in unstimulated cells to 0.85 h in cells stimulated with lymphotoxin. In

addition, stimulators of protein kinase C, including phorbol esters and teleocidin, enhanced

accumulation of IL-6 mRNA. Cycloheximide (CHX), inhibitor of protein synthesis, also

markedly increased levels of IL-6 mRNA. Both CHX and activators of protein kinase C

increased by greater than 16-fold the stability of IL-6 mRNA. Further, dose-response studies

showed that sodium fluoride (NaF), activator of G-binding proteins, and ouabain, inhibitor of

Na+/H+ pump, increased levels of IL-6 mRNA. NaF stimulated IL-6 mRNA levels

independent of protein kinase C activity. These results suggest that stimulators of several

pathways of signal transduction increase levels of IL-6 mRNA and posttranscriptional

stabilization is, in part, the mechanism that many of these signals, including lymphotoxin,

use to increase […]

Research Article

(2)

Role of Lymphotoxin

in

Expression

of

Interleukin 6 in Human

Fibroblasts

Stimulation and Regulation

Makoto Akashi, ArthurH.Loussararian, Daniel C.Adelman,* Masaki Saito,t and H. Phillip Koeffier

DivisionofHematology/Oncology, and*DivisionofClinicalImmunology/Allergy, DepartmentofMedicine, UniversityofCaliforniaLosAngeles,LosAngeles, California 90024; and

$Division

ofHemopoiesis, InstituteofHematology, Jichi Medical School, Tochigi-ken 329-04, Japan

Abstract

IL-6 is

a

cytokine

with a number of biological functions,

in-cluding

stimulation of

immunoglobulin synthesis

and prolifera-tion

of

early

hematopoietic

stem cells. We showed that

lym-photoxin stimulated

accumulation

of IL-6 mRNA in human

fibroblasts

(W138)

in a

dose-responsive fashion;

tumor

ne-crosis

factor-alpha (TNF-a)

wasabout threefoldmorepotent than

lymphotoxin.

Further

experiments suggested

that stimu-lation by

lymphotoxin

was independent

of

protein

kinase

C

activity,

did not

require

new

protein

synthesis,

andwas atleast in part a result

of increased stabilization of

IL-6 mRNA.

t1/2

of the

IL-6

transcripts increased from 03

hin

unstimulated

cells to

0.85

h in cells

stimulated

with

lymphotoxin.

In

addition,

stimulators of

protein kinase C, including phorbol

esters and

teleocidin,

enhanced

accumulation

of IL-6 mRNA.

Cyclohexi-mide

(CHX), inhibitor of protein synthesis, also markedly

in-creased levels of

IL-6

mRNA. Both CHX and activators of protein kinase

C

increased by >

16-fold

the

stability

ofIL-6 mRNA. Further, dose-response studies showed that sodium

fluoride

(NaF),

activator

of

G-binding

proteins,

and

ouabain,

inhibitor of

Na+/H+

pump, increased levels

of IL-6

mRNA. NaFstimulated

IL-6

mRNAlevels

independent of protein

ki-nase Cactivity. These results suggest that stimulators of sev-eral

pathways of signal transduction increase

levels

of IL-6

mRNA and

posttranscriptional stabilization is,

in part, the

mechanism

thatmany of these

signals, including lymphotoxin,

use to increase levels

of

IL-6 RNA.

(J. Clin.

Invest.

1990.

85:121-129.)

messenger RNA -

posttranscriptional

stabili-zation

Introduction

IL-6

is

a 2

l-kD glycoprotein containing 184 amino acid

resi-dues derived from aprecursor

peptide

of 212 amino acids. The amino-terminal of IL-6 shares homology with

granulocyte-colony stimulating factor

(G-CSF)'

(1).

IL-6

contains five

exons

(2)

and

is located

on chromosome 7

(3).

This

cytokine

was

previously

known as

IFN-#f-2,

26K

protein,

B-cell

stimu-lating factor-2, hybridoma growth factor,

or

hepatocyte-stimu-lating factor

(4). IL-6appears to have a variety of

functions,

Addressreprint requests to Dr. M. Akashi, UCLA Division of Hema-tology/Oncology, 11-240 Factor Building, Los Angeles, CA 90024-1678.

Receivedforpublication 27 March 1989 and inrevisedform24 July 1989.

including stimulation

of

proliferation

of

early hematopoietic

stem

cells

and

megakaryocytes,

induction of differentiation of B

lymphocytes, and stimulation

of the liverto

produce

acute

phase-reactive proteins.

Tumor

necrosis factor-alpha (TNF-a),

also known as

TNF,

is

produced

predominately

by

myeloid

cells, especially macro-phages (5, 6).

Lymphotoxin,

also known as

TNF-,6,

is

synthe-sized

predominately by activated lymphocytes (5, 7, 8).

Lym-photoxin

has structural

similarities

to TNF

and both

are

cyto-lytic

or

cytostatic for transformed

cells

(9, 10). Previously,

Kohase

et al.

(11) reported

that TNF-a and several other agents,

including

a

phorbol

ester, stimulated normal

fibro-blasts

to

synthesize IL-6, and their

data

suggested

that the

stimulation

was

probably

mediated

through

activation of the

protein kinase C pathway (12).

However,

little is

known

ofthe

role that lymphotoxin plays in modulation of IL-6 expression.

In

this

study, we show

that

stimulators of several pathways of

signal transduction

elevate

levels

of

IL-6 mRNA and

stabiliza-tion

of

these

transcripts is, in

part,

the

mechanism

that many

of

these

signals, including lymphotoxin,

use toincrease levels

of

IL-6 mRNA.

Methods

Reagents. Recombinant human TNF-a (5.6 X 107 U/mg protein) and recombinant lymphotoxin (2 X 108 U/mg protein) were generously suppliedbyM.Shephard, Genentech Inc. (South San Francisco, CA); both wereexpressed in Escherichiacoli.Both contained lowlevelsof endotoxin (TNF-apreparation,<0.03 EU/mgprotein; lymphotoxin preparation, 3.13 EU/mg protein). Specific activity of each was as-sayed twice;once atGenentechand thenatUCLA by examining their cytolytic activityonactinomycin-D-treated L929 fibroblasts (6).

Re-sults were nearly identical. 12-0-tetradecanoylphorbol 13-acetate (TPA), 4-0-methyl TPA, phorbol 12,13-didecanoate (PDD), 4a-PDD, and phorbol 12,13-dibutyrate (PDB) were purchased from Sigma Chemical Co. (St. Louis, MO), dissolved in acetone, and stored at -20'C until use.Allother reagents were also ofthe best grade available from Sigma Chemical Co.

Cells and cell culture. Normal human embryonic lung fibroblast (W138,obtained from the American Type Tissue Culture Collection) anda human lung adenocarcinoma cell line (Lu-CSF-l) (13) were cultured inamedium (Flow Laboratories, Inc., McLean, VA) supple-mented with 10% FCS in a humidified atmosphere containing 5% CO2. Cellswere harvestedby treatment with 0.05% trypsin, 0.02%

EDTA(wt/vol) in PBS. Conditioned media (CM) from confluent

cul-turescontainingWI38 fibroblasts and either TNF-a orlymphotoxin wereprepared by centrifuging the supernatants at 1,000 g for 10min,

1.Abbreviations used inthis paper: CHX, cycloheximide; CM, condi-tioned media; G-CSF,granulocyte-colony stimulating factor; GM-CSF,granulocyte monocyte-colony stimulating factor, PDB, phorbol 12,13-dibutyrate; PDD, phorbol 12,13-decanoate; TNF-a and

-#,

tumornecrosisfactor-aand

-#,B

respectively; TPA,

12--tetradecanoyl-phorbol 13-acetate.

J.Clin.Invest.

(3)

puttingthem through filters (0.22Mm;Millipore Corp., Bedford, MA), andstoringat-20'Cuntil use.

Measurement ofIL-6bioactivity. IL-6 levelswere measuredusing the IL-6-dependent murine hybridoma clone, MH60.BSF2, in a

bioassay(14). Thesehybridomacells areabsolutely dependent on IL-6; IL-la and

fl,

IL-2, IL-3,IIA,IL-5,IFN-aor-y,and G-CSF do not

induce their growth (12). 1 X 104 cells/well were cultured in RPMI-1640medium containing10% FCS (lot 5121103; Irvine

Scien-tific, SantaAna,CA), 50 U/ml penicillin, 50Mg/mlstreptomycin, 2

mM L-glutamine,and 5 X 10-5 M 2-mercaptoethanol with test

sam-plesorvarying concentrations ofrecombinant IL-6 (5 X 106 U/mg

protein;generousgift ofCetus Corp.,Emeryville, CA [14]). Each sam-ple and standard were studied at different dilutions in triplicate, round-bottom, 200 Ml microtiterwells(NunclonO Delta, InterMed Nunc, Roskilde, Denmark) for 48 h in a 5%CO2humidified incubator at370C,pulsedwith 2MCi

[3H]thymidine

(20.0 Ci/mmol; New En-gland Nuclear, Boston, MA) for the last 6 h, harvested onto glass filter paper, and prepared for liquid scintillation. Radioactivity was

mea-suredon ascintillationcounter (Beckman Instruments, Inc., Fuller-ton, CA), and data were analyzed and graphed on a Macintosh II

personalcomputerusingStatWorks'& (Version 1.0,Datametrics,Inc.,

Philadelphia,PA) and Cricket Graph (Version 1.2, Cricket Software,

Malvern, PA)software.

Studies ofantibody neutralization of IL-6 activity in CM used a

heterogenous rabbit antisera raised against recombinant IL-6

(Gen-zymeCorp., Boston, MA).It wasusedat 1 Mg/ml. Aliquots of fibro-blasts CMwerepreincubatedfor 2hat37°C with IL-6 antibody and then addedtotheMH60.BSF2 cells.

DNAprobes. Human IL-6 cDNA (1.2 kb, XhoI)was derived from plasmid pXM (generous gift of S. Clark, Genetics Institute, Cam-bridge, MA).

_fl-actin

DNAprobe (0.7 kb, EcoRI-BamHI)wasfrom

pHF-#

A-3'atplasmid (15). These probeswere32P-labeledby random priming method (16). The specific activitywas5-8X

101 cpm/Mg.

Isolation andblottingofRNA.ForcytoplasmicRNA, WI38 cells

weresuspended in hypotonic buffer containing 10 mM Tris-HCl (pH 7.4), 1 mM KC1, 3 mM MgCI2, and werelysed with 0.3% NP-40. Cytoplasmic RNA was extracted by phenol/chloroform method as previously described (17). After denaturationat65°C,RNAwas elec-trophoresed inanagarose-formaldehyde gel (1%)andtransferredto a nylon-membrane filter (BiotransO, ICN,Irvine, CA) (18). Hybridiza-tion with labeledprobewasfor 16-24 hat42°Cin 50%formamide,2X SSC (1X = 150 mM NaCl, 15 mM sodiumcitrate), 5XDenhardt's solution, 0.1% SDS, 10% dextran sulfate, and 100 Mg/mlsalmon

sperm. Filterswerewashed to astringencyof 0.1X SSC, 65°Cand exposedtoXARfilm(EastmanKodakCo., Rochester, NY).

Autora-diogramsweredevelopedatdifferentexposures.

Blotswereusually sequentially hybridized with 32P-labeled IL-6 and,B-actinDNA.The actin band ofhybridizationwasusedtohelp confirm that similaramountsofRNAwereaddedtoeach lane. Modu-lation in levels of IL-6RNAwerequantified byinitialstandardization

totheamountof

f-actin-specific

transcripts. Therelativedensityof

f3-actin-specific

andIL-6-specifictranscriptsin thedifferent laneswas

firstdetermined bylaserdensitometryusing multipleexposuresof the blot,and theratioofIL-6/,B-actinin thecontrollanewas

assigned

tobe the baselinelevel.Thefold stimulationintheexperimentallaneswas

calculated bymultiplyingthe ratio ofdensityof

(IL-6/1B-actin)

tran-scripts by thereciprocalof the ratio of the baselinelevel.

[3H]Uridineand

[35S]methionine

incorporation. Fibroblastswere exposedto protein kinaseC inhibitors orcycloheximide (CHX)in culturedishes(Falcon",BectonDickinsonLabware,Oxnard,CA)in triplicateperexperimental pointfor2h.Cellswerepulsedwith2MCi of[3H]uridine (43 Ci/mmolspact)or4MCi of

[35S~methionine

(200

mCi/mM) for 1 hat37°C, washed twice inPBS,precipitated in 5% TCA in 30 mMNa2HPO4at4°Cfor 1h,filteredonto aglass micro-fiber membrane(Whatman Inc.,Clifton, NJ; GF/F),washed in 3% TCA(30 mMNa2HPO4),andheatedat80°CforI h.Eachsamplewas countedbyliquidscintillation. Resultswerecomparedwith those of

untreated cells.

Protein kinaseCassay. The activityofproteinkinase C was as-sayedasdescribedpreviously(19).Briefly,soluble(cytosol)and solubi-lizedparticulate(membrane) fractions ofcellswereprepared, and both fractionswereassayed for protein kinase Cactivity usinghistoneH I as

substrate. Theamountsofproteinkinase C in thecytosolic (22Mtg)and solubilized membrane(5

Mg)

fractionsweredetermined in the reaction mixtures (0.2 ml) containing 25 mM Tris-HC, pH (7.5), 10 mM MgCl2,5

MM

['y-32P]ATP,

and40

Mg

ofhistone.

Results

Lymphotoxin induction of

IL-6

protein

in

fibroblasts.

Lym-photoxin

was

added

to

subconfluent cultures of W138

lung

fibroblasts

for

4

d; CM

were

harvested for IL-6

activity (Fig. 1).

Inthe

absence of

lymphotoxin,

fibroblasts

constitutively

pro-duced IL-6

protein. Lymphotoxin

enhanced

production

of

IL-6 in

a

dose-dependent

manner.

104 U/ml

lymphotoxin

stimulated

approximately

five times

greater

concentration of

IL-6

than

was

constitutively produced.

The

activity

of IL-6 in

the CM cells cultured with

103 U/ml

lymphotoxin

was

almost

neutralized by

IL-6

antibody (78±1%).

Additional control

ex-periments

showed

that

fresh

culture

media

containing 103

or

104 U/ml

lymphotoxin

inhibited

[3H]Tdr incorporation

in the

IL-6

responder

cells

(MH60.BSF2) by

<5%

of untreated

con-trol cultures

(data

not

shown).

70

-60

-E

-J

D c

50'-

40'-

30'-

20'-

10'-10

POSSES

E I-

--.20

CF FFCC £ C

0 0 0 0 0

U U U C C

Figure

1.IL-6

_production

after exposure of fibroblaststo

lympho-toxin

(LT).

Human

lung

fibroblastswereculturedwith

_0-104

U/ml

lymphotoxin

for4

_d;

conditioned media

(CM)

wereharvested and

added

(50%

vol/vol)

to

_{IL-6-dependent}

murine

_hybridoma

clone,

MH60.BSF2asdescribed in Methods. Resultsrepresentmeanand

(4)

Lymphotoxin

TNFa

UJ/mil

C

10 1o2103104

1O 102,03

104 1.3kb-'-

*

0 *

*

IL-6

1

3

4

3

8 10 12

Fold

Stimulation

2.1 kb-

) -

actin

Figure 2. Dose-dependent effect oflymphotoxin and TNF-a on levelsofIL-6mRNA in human lung fibroblasts. Fibroblasts were cultured with eitherlymphotoxin or TNF-a for 8 h. Cytoplasmic RNA(15 lane)was par andanalyzed by formaldehyde-aga-rose gelelectrophoresisandtransferred to a nylon membrane as de-scribed in Methods. Hybridization waswith32P-labeledIL-6 cDNA

(1.3 kb; top)

and

#l-actin

DNA (2.1 kb; bottom). Fold stimulation

of levelof

IL-6

mRNA as compared with levels in untreated cells was equalized for levels of

fl-actin

(seeMethods).

Dose-dependent

effect

oflymphotoxin and

TNF-a on

levels

of

IL-6 mRNA.

Fibroblasts

constitutively

contained a

low

concentration

of mRNA

coding

for IL-6

(Fig. 2).

After expo-sure to

10-104 U/ml

of either

lymphotoxin

or TNF-a for8h,

levels

of IL-6

mRNA

increased

in a

dose-dependent

manner. TNF-a was more potent

than

lymphotoxin.

10

U/ml

TNF-a

stimulated

an

equivalent

accumulation of IL-6 mRNA as

102_103 U/ml lymphotoxin;

and at

maximally

stimulatory

A

Teleocidin

/lg/mI

C

5

50 concentrations,

TNF-awas threefoldmore potent than

lym-photoxin. Time-response experiments (0-8

h)

found that both TNF-a

(3,000 U/ml)

and

lymphotoxin (3,000 U/ml)

stimu-. lated accumulation of IL-6 mRNA with a similar

_rapidity,

with maximal stimulation

achieved within 1 h

(data

not

shown).

Aswasfound in the

dose-response studies,

TNF-awas

fourfold more potent than

lymphotoxin

at each time

point

(data

not

shown).

Effect

ofteleocidin and

various

derivatives

ofphorbol

ester on expression

of

IL-6 mRNA.

Teleocidin

is a

nonphorbol

tumor promoter which can activate

protein

kinase C

(20).

Treatment of cells with 5 or50

jg/ml

teleocidin for 2 h in-creased accumulation of IL-6 mRNA about fourfold

com-pared with untreated cells

(Fig.

3

A). TPA, PDD,

andPDBare

phorbol esters thatare potentactivators of protein kinase

C;

their

derivatives, 4-0-methyl

TPA and 4-a-PDDareunableto activate protein kinase C

(21).

Fibroblasts were

exposed

to each compound (50 nM) for 2 h and levels of IL-6 mRNA measured.Potency of the phorbol ester tostimulate accumu-lation ofIL-6 mRNA in fibroblasts

paralleled

their known abilitiestoactivate

protein

kinase C

(Fig.

3

B).

Effect

of

inactivation

of

protein kinase

C

on

expression

of

IL-6mRNA

stimulated

by either NaF

or

lymphotoxin. NaF is

anactivator of several G-binding proteins (22). To determine whether NaF stimulates IL-6 mRNA accumulation, fibro-blastswereculturedwith NaFatdifferent concentrations for 4 h. 5mM NaF stimulateda28-fold increased accumulation of IL-6 mRNA as compared with untreated control cells

(Fig.

4A).

Cells exposed for prolonged durationstoTPAreduce their protein kinase C activity, thus making them resistant to re-peated exposuretoTPA(23). We showed that treatment with 100 nM TPA for 24 h decreased protein kinase C activity

by

>80% and decreased cell membrane-binding of [3H]PDB

by

>90%, andreexposure toTPA didnotstimulatethe

activity

ofprotein kinase C (data not shown). Fibroblasts incubated with TPA (50 nM) for 4 h contained markedly increased levels of IL-6mRNA compared with untreated cells (Fig. 4 B).

Pro-B

Phorbol ester

IC

1

2

3

4

5

1.3

kb

I

Fold Stimulation

1 4 4

2.1 kb-

_

*

1 12 2 10

2

10

Figure 3.Effect of teleocidin (A) and var-ious derivatives of phorbol ester (B) on ex-pression ofIL-6mRNA infibroblasts. (A) Cells culture with either 5 or 50 nM teleo-IL - 6 cidin for2 h. Lane C, untreated control.

(B) Cellswereexposed to eachcompound (50nM) for 2 h: lane 1, TPA; lane 2, 4-0-methyl TPA; lane 3, PDD; lane 4, 4-a-PDD; lane 5, PDB. Lane C, untreated con-trol. Northern blot analysis of mRNA was performedby blotting cytoplasmic RNA

(15 jAg/lane).

The 1.3- and 2.1-kb

hybridiz-2-

ac

t

i

n

ing

bandsareconsistent with mRNA

(5)

D =: CL Xd

B

NaF

[ I

C

Q5

1

5

U) U)

-C

cr

LL-H-

Zz

+ +

- U) -- U) th L- U) L. L- h- r_

n) N rON < <: LL <;

a a. aOC

HHFZ

.0

.IL-6

Fold Stimulation

1

2

28 2.1kb-

a *

a

p-actin

L _I! I L a L__ J

%

Accumulation

100 90

1OO

41 100 326 100 126

2.1

kb-Figure 4. Effect of prolongedexposureofaphorbolesteronexpression of IL-6 mRNA stimulated byNaForlymphotoxin.(A)Fibroblastswere

cultured withdifferentconcentrations (0.5-5 mM) of NaF (4 h). (B) Cellswerepretreated with TPA (100nM,24h), washed three times, and

exposedtoeither TPA(50 nM), NaF (5 mM),orlymphotoxin (4,000 U/ml) for 3 h. As positive control, cellswereculturedonlywithTPA (50

nM, 3 h), NaF (5 mM, 3 h)orlymphotoxin (4,000 U/ml, 3 h). Cytoplasmic RNAwasextracted and each blot(I5

_gg/lane)

wassequentially

hy-bridized with IL-6 cDNA(1.3 kb) and

fl-actin

DNA(2.1 kb). Forpercentaccumulation, each positive control (3 hexposuretoTPA, NaF,

lym-photoxin)wascompared with cells pretreated withTPA(24 h) and then cultured (3 h) with TPA, NaF,orlymphotoxin.

longedexposure(24h)of cellstoTPA didnotstimulate their accumulation of IL-6mRNA;reexposureofthese cellstoTPA for 4 hhad60% lowerlevelsof IL-6 mRNAascompared with

those cells cultured with TPA for 4 h alone. In contrast, the enhanced levels of IL-6 mRNA mediatedbyNaFor lympho-toxinwerenotblockedbypreexposureof thecellstoTPAfor 24 h(Fig.4B).Thecombinationof 24-hexposuretoTPAplus

anadditional 4 h of TPA and NaF stimulated the accumula-tionof IL-6 mRNAslightlymorethan NaF alone. Thereason

forthisenhanced stimulation isnotclear. Theseexperiments

were repeated three times with similar results (data not shown).

Effect ofouabain

onexpression

of

IL-6 mRNA. Fibroblasts wereculturedwithouabain(inhibitorof

Na+/H+

pump;24); cytoplasmicRNAwasextracted andanalyzedforIL-6 mRNA levels.Exposuretoouabain markedlyincreasedaccumulation ofIL-6mRNAinadose-dependentmanner(Fig. 5).

Time-re-sponseexperimentsshowed that within 4h,ouabainincreased IL-6 mRNA levels by 20-fold, and a 100-fold increase oc-cuffedat24 h(datanotshown).

Effect of lymphotoxin

onexpression

of

IL-6 mRNA in

the

absence

ofprotein synthesis

in

fibroblasts. Fibroblasts cultured

with20 sg/ml CHX (inhibitor of protein synthesis) decreased protein synthesis by 93% ascompared with untreated cells. Cells culturedinitiallywith CHX(20.0

,g/ml)

andfollowedby

theadditionoflymphotoxin hada98-fold enhanced

accumu-lation of IL-6mRNAascompared with untreated control cells (Fig. 6). This enhancement wasgreater than those observed when the cellswerecultured with either lymphotoxin (fivefold) orCHX(59-fold)alone.

Stability ofsteady-state

IL-6

mRNA

in

fibroblasts exposed

to either

lymphotoxin, TNF, TPA,

or CHX. To

examine for

posttranscriptional regulation ofexpression of IL-6 mRNA, untreated fibroblastsorthoseexposedtolymphotoxin (4,000 U/ml),TNF-a(1,000 U/ml),TPA(50 nM),orCHX(5

Ag/ml)

for 4 h were also cultured withactinomycin D for an addi-tional 0.5-4.0 h. Samplesweresequentiallyharvested and ex-amined for levels of IL-6 mRNA. t1/2 ofsteady state IL-6 mRNA in unstimulated fibroblastswas 0.3 h (Fig. 7).

tI/2

of IL-6 mRNAwas0.85 hafter culture with eitherlymphotoxin

orTNF-a(Fig.7) andwas >4 hafterexposuretoeitherTPA

orCHX(Fig. 8). Similar resultswereobservedinthreemore

experiments.

Discussion

We have examined levels of IL-6 mRNAin fibroblasts after theirexposuretoeitherlymphotoxinorstimulatorsofseveral

signal pathways. Lymphotoxinisaglycoprotein produced by

activated lymphocytes (7, 8, 25-27);itcaninhibit thegrowth

A

In

U)

rO

rnr

CUj

-+

U)C

1.3 kb

- IL-6

1.3 kb-A8-

actin

(6)

Lu-CSF-1

Ouabain

C

1o&8107iCT6ic5j64

i(

1.3

1.3 kb-Fold Stimulation

1

1 10 18

21 Fold

Stimulation 5

59 98 1

2.1

kb-Figure5. Effect of ouabainonexpressionof IL-6 mRNA. Fibroblasts

wereexposedtodifferent concentrations of ouabain for 24 h (10-8-lO- M). Northern blot analysis of IL-6 and

fl-actin

mRNAwere

per-formedby blotting cytoplasmic RNA (15

gg/lane)

from cellsas de-scribedinMethods. Fold-stimulationwascalculatedasstated in Methods.Lu-CSF-I mRNA (lung adenocarcinoma cell line)was

usedas apositive control (13). Fold stimulation of IL-6 mRNA levelsdeterminedasstated in Methods. The minor bandsseenat

1.3 kb inthe

fl-actin

panelrepresentradioactivity thatwasnot

re-movedafter hybridization with the IL-6 probe.

ofavariety of cells transformedbychemicals orviruses (28, 29). Manyofthe actions oflymphotoxinaresimilaror identi-caltothose observed with TNF-a. Lymphotoxin andTNF-a

probablyshare thesamecell surface receptors(30). Previously,

wehave observed thatlymphotoxin stimulates normal fibro-blasts to produce several hematopoietic CSFs (31). In this study, we have found that normal lung fibroblasts constitu-tively producedIL-6 mRNA. Afterexposuretolymphotoxin, levels of IL-6 mRNAincreasedinthese cellsmorethan four-fold. TNF-awasaboutthree timesmorepotent than lympho-toxinatequivalentconcentrations of thesecytokines.

Proteinkinase C is involvedinsignaltransductionby

cou-pling receptor-mediated inositol phospholipidturnoverwitha

varietyof cellular functions(32). We have observed that both phorbol diesters and nonphorbol compounds that activate proteinkinase C stimulated accumulation of IL-6 mRNA in the fibroblasts. In contrast,phorbolestersincapableof activat-ing protein kinase Cwereunabletostimulateincreased

con-centrations of IL-6 mRNA. Inaddition,wetookadvantageof the fact thatprolongedexposureof cellstoTPAleadsto inac-tivation ofprotein kinase C(33, 34). Prolonged (24 h)

expo-sure to TPA partially blocked accumulation of IL-6 mRNA afterreexposure ofcellstoTPA. However under these same

conditions, accumulation of IL-6 mRNA was not blocked after reexposure of the TPA-treated cells to lymphotoxin. Zhang et al. reported that TNF and IL- stimulated IL-6

Figure 6.Ability of lymphotoxintostimulate accumulation ofIL-6 mRNA infibroblasts in absence of protein synthesis. Cellswere ei-ther treatedwith lymphotoxin (2.5 h, 3,000 U/ml), (CHX, 2.5 h, 20

,ug/ml)orpretreated with CHX(0.5 h) and then cultured with CHX and lymphotoxin (2.0 h). Analysiswasperformed by blotting

cyto-plasmic RNA(15 ug/lane) and hybridizing with 32P-labeled IL-6 (1.3 kb)orfl-actin(2.1 kb) DNA probesasdescribed in Methods. Fold

stimulation of-bands of IL-6mRNAwasdeterminedasdescribed in Methods. The bandsseenat - 1.3 kb in thepanel offl-actin

repre-sentradioactivity thatwasnotremoved afterprevious hybridization with the IL-6 probe.

mRNAlevels byaprotein kinase C-independent mechanism in human fibroblasts using similar methods (35). Taken

to-gether, these results also suggest that lymphotoxin probably doesnotmediate accumulation of IL-6 mRNAthrough

pro-teinkinase C. Anotherstudy using putative inhibitorsof

pro-tein kinase C suggested TNF-a stimulated accumulation of IL-6 mRNA through the protein kinase C pathway (12). We alsohave found that thesameinhibitors(H-7andH-8,50 mM for 2h)could block the stimulationmediatedbyboth TNF-a andlymphotoxin;but theseagentsalso inhibited the stimula-tion mediatedby NaF, CHX,andouabain,and decreased total RNAsynthesis by 70% (data notshown). In view of thevery short 1/2of IL-6 mRNA,webelieve that theabilityof H-7to

inhibit the accumulation of IL-6 mRNA mediated by lym-photoxin is probablyanonspecific effect.

Arecentstudy suggested thatG-binding proteinsmay be

involved in transducing the signal mediated by TNF after bindingtoits receptorsonleukemiacells (36). Inthisregard, wehave found thatNaFstimulated the accumulation of IL-6 mRNA. F- ion in combinationwith Al3+ (usuallypresentin

trace amountsintissueculture) binds and activates G-binding proteins. Preexposure ofthe fibroblasts to pertussis toxin (10-1,000 ng/mlfor 4-24h)wasunabletoblock the action of

1-+ 0

~~jOOC~~~h

2.1 kb-IIL-

6 ,B-actin

,8-act

in

I L- 6

.l''T

*..

FT

.-

-F

,,j.

Iz.

(7)

A

Untreated Hours 0 0.5 1 2

B

Lymphotoxin C 0 0.5 1 2

C

TNFa C 0 0.5 1 2

kb

1.3 ON*t t1

',W'";'i;:'.:'

2.1-D

1--W

GD

9' 100

50

10

s

1

Hours

2

IL-6

Figures7and 8. Stability of steady-stateIL-6 mRNA in humanfibroblast exposed to lym-photoxin,TNF-a, TPA,orCHX. Untreated cells(Fig.7A)orthoseexposed to lympho-toxin (4,000 U/ml,Fig.7B);TNF-a(1,000 U/ml, Fig.7C);CHX (5

jug/ml,

Fig. 8 A)or

TPA(50 nM, Fig. 8 B) for4h werethen also cultured with actinomycinD(5

Ag/ml)

for 0.5-4.0 h.Cytoplasmic RNA (30

pg/lane

in

untreatedcells and 15

pg/lane

in lympho-toxin,TNF,CHX,orTPA-treatedcells)was

extractedandanalyzed byRNAblottingas de-scribed in Methods. Blots were sequentially hybridized with

32P-labeled

IL-6

cDNA(1.3 kbbands, toppanels) and with

jlactin

DNA (2.1 kb bands, bottompanels).Intensity of hy-bridizationwasdetermined by densitometory of several different exposures of the autoradio-grams. PlotsofRNAstabilityareshown on Fig.7DandFig. 8C.Untreated cellsof

each

experiment

wereassumedtohave100% activ-ity.

NaF

(data

not

shown). Several G-binding proteins

are not

ri-bosylated and therefore inactivated

by

pertussis

toxin

(37).-NaF may

modulate levels of IL-6

mRNA

in

fibroblasts

through

a

G-binding protein pathway

that is insensitive

to

pertussis

toxin.

In

further

experiments,

our

results

suggest

that

the increase of

levels

of IL-6

mRNA

mediated by

NaF

proba-bly

does not

require protein

kinase C because

partial

inactiva-tion of this

enzyme

did

not

blunt

the response of NaF.

We

observed

that

ouabain markedly increased levels of

IL-6

mRNA

in

a

dose-dependent

manner.

Ouabain blocks the

Na+/K+

pump,

causing

a

rise

of intracellular Na+

(24).

In-creased intracellular

Na+ levels

can

increase the

cytoplasmic

Cal' concentration.

Prior

studies showed

thata

Ca"

iono-phore

can enhance

levels of IL-6

mRNA

(12).

Further studies

are inprogress to

determine

the

importance

of Na+

and/or

Ca++ cellular fluxes

in

mediating

the accumulation of IL-6

mRNA.

The

IL-6

mRNA

level increased in the fibroblasts

when new

protein synthesis

was

nearly stopped by

CHX.

Seghal

et al. also showed

that

exposure

of human

diploid

fibroblasts

to

CHX increased their

concentrations

of

IL-6 mRNA

(9).

In

addition,

we

have

found that

lymphotoxin

was

able

tocause

the

accumulation of IL-6

mRNA

in the absence of

new

protein

synthesis.

To pursue

further

how

lymphotoxin stimulated

the

in-crease

in levels of IL-6 mRNA,

we

studied

stability

of IL-6

mRNA. In

untreated

fibroblasts,

the

_t,/2

of IL-6

mRNAwas

0.3

h.

Exposure of the cells

to

either CHX

or TPA

markedly

stabilized

the message,

with little

decrease in

levels

at 4 h

of

exposure. IL-6 mRNA

is

a

transiently

expressed

gene

having

an

AU-rich 3'

region,

which

may beatarget

for

short-acting

RNAses.

Studies of Shaw and

Kamen

(38) suggested that

TPA

and

CHX

stabilized

another

cytokine

(GM-CSF) by altering

the response

of

the cell to

this AU-rich

region

of

the mRNA. TNF-a

and lymphotoxin

also

consistently increased the

stabil-ity of IL-6 mRNA,

both

increasing the

t1/2

of IL-6

mRNA to

-

0.85

h as

compared with

the

untreated, control fibroblasts.

Further studies

are now

determining if the stabilization of

mRNA

by

lymphotoxin requires the

AU

canonical

sequences

in the

mRNA.

We

previously

have

found that

TNF-a

can increaselevels

of

G- and

GM-CSF

mRNA

by

stabilization

of

these RNAs

(13).

A

posttranscriptional regulation by

various

W,, *,z,. iw

40:

a

ML, _-actin

(8)

A

CHX

C

0

0.5

1

2

4 kb

1.3-

2.1-

C--I-w

4.

1;

*W

B

_TPA

C

0

0.5

1

2 4 Hours

O OM to

_

IL-6

,B

-actin

Figures 7 and 8 (Continued)

signals, including TNF-a and lymphotoxin,maybea common

mechanismfor rapid and sensitive changeinlevels of IL-6and other cytokines. Because both TNF-a and lymphotoxin equally stabilized the IL-6 mRNA,anotherexplanation is

re-quired to account for the threefold increased potency of TNF-aascompared with lymphotoxininstimulating levelsof IL-6 mRNA. Experimentstomeasureratesof IL-6

transcrip-tion

in fibroblasts exposed to lymphotoxin and TNF-a are nowinprogress.

Fibroblasts constitute a major element ofbone marrow

stroma,as wellassubmucosal andsubcutaneous tissue. Our datasuggestthatinvivo, lymphotoxin produced byactivated lymphocytescanstimulate these mesenchymalcellsto synthe-sizeIL-6. Undernormalconditions, IL-6canhavenumerous

functions, including the stimulation of the proliferation of early hematopoieticstemcells(39)and the differentiation of B lymphocytes (40). Lymphotoxin and IL-6 may also be in-volved in pathologic conditions. For example, affected joint

spacesinrheumatoidarthritis often containhighlevels ofIL-6,

lymphotoxin, TNF-a, and IL-1 (41-43). Thesefactorsatleast inpartarecapable of causing bone resorption,aswellas

colla-genand cartilagedestruction(42-45).

Furthermore,

theBand Tlymphocytes, macrophages, and synovial cellswithin these joints may inappropriately develop a paracrine network through productionof thesecytokines,helpingtomaintainthe chronic

inflammatory

rheumatoidjoint.Fundamental under-standingofcytokinesand their interactionsmayleadtonovel approachestoseveral diseases.

Acknowledgments

The authors thank Dr. A. Saxon for helpful discussions; Dr. J. F. Kuo (Emory University, Atlanta) for generously measuring protein kinase

Cactivity; Mr.T.Yen forexperttechnicalassistance; and Elisa Weiss, Margery Goldberg, and Elaine Epstein forexpertsecretarial help.

This workwassupported inpartby U. S. Public Health Service

grantsand the 4E Leukemia Memorial Fund inmemoryofMarilyn Levine,ErwinEpstein, Goldie Berman, Harold Bass, and in honor of RoselleLewis.

Hours

T

'I

_,. ..-

"..

s.s.1

VW

4l

(9)

References

1.Hirano, T.,K.Yasukawa, H. Harada,T.Taga, Y. Watanabe, T. Matsuda,S.Kashiwamura,K.Nakajima,K.Koyama, A. Iwamatsu, S. Tsunasawa,F.Sakiyama,H.Matsui,Y.Takahara,T.Taniguchi, and T.Kishimoto. 1986.ComplementaryDNAforanovelhuman inter-leukin(BSF-2)thatinducesBlymphocytestoproduce immunoglobu-lin. Nature(Lond.).324:73-76.

2.Yasukawa, K., T.Hirano,Y.Watanabe, K. Muratani, T. Ma-tsuda,S.Nakai,andT.Kishimoto. 1987. Structure and expression of human B cell stimulatory factor-2 (BSF2/IL-6) gene. EMBO (Eur. Mol.Biol.

Organ.)

J.6:2939-2945.

3.Sehgal,P.B.,A.Zilberstein,M. R.Ruggieri,and F. H. Ruddle. 1986. Human chromosome7carries the beta 2 interferon gene. Proc. Natl. Acad. Sci. USA. 83:5219-5222.

4.

Gauldie,

J.,C. Richards,D.Harnish,P.Landorp,and H. Bau-mann. 1987.Interferon beta2/B-cell stimulatoryfactor type 2 shares

identity

with monocytes-derived hepatocyte-stimulatingfactor and regulatesthe majoracutephase proteinresponse in livercells. Proc. Nail.Acad.Sci. USA. 84:7251-7255.

5.Goeddel,D.V.,B.B.Aggarwal,P. W.Gray,D. W. Leung, G. E. Nedwin, M. A.Palladino,J.S. Pastton, D. Pennica, H. M.Shepard,

B.J.Sugarman,and G.H.W.Wang. 1986. Tumor necrosis factors: gene structure and biological activities. ColdSpringHarbor Symp.

Quant.

Biol. 51:597-609.

6.

Pennica,

D.,G. E. Nedwin, J.S. Hayflick, P. H. Seeburg,R.

Derynck, M. A. Palladino, W. J. Kohr,B. B. Aggarwal, and D. V.

Goeddel. 1984.Humannecrosis factor: precursor structure,expression andhomologyto

lymphotoxin.

Nature

(Lond.).

312:724-729.

7.Aggarwal, B.B., B. Moffat,and R. N.Harkins. 1984.Human

lymphotoxin:

production bya

lymphoblastoid

cellline, purification, andinitial characterization. J.Biol. Chem. 259:686-691.

8.

Stone-Wolff,

D.S.,Y. K.Yip,H.C.Kelker,J.Le, D.

Henrik-sen-DeStefano,

B. Y.

Rubin,

E.

Rinderknecht,

B. B.Aggarwal,and J. Vilcek. 1984. Interrelationships of humaninterferon-gamma with

lymphotoxin

and monocytecytotoxin.J.Exp.Med. 159:828-843. 9.

Gray,

P.W.,B. B.Aggarwal,C.V.Benton,T.S.

Bringman,

W.J.

Henzel,

J. A.Jarrett,D. W.Leung,B.

Moffat,

P.

Svedersky,

L. P.Ng, M.A.

Palladino,

and G. D.Nedwin. 1984.

Cloning

and

expressing

cDNAfor human

lymphotoxin,

a

lymphokine

withtumornecrosis

activity.

Nature

(Lond.).

312:721-724.

10.Sugarman,B.J.,B. B.Aggarwal,P. W.Hass,I. S.

Figari,

M. A.

Palladino,

and H. M. Shepard. 1985. Recombinant humantumor necrosis

factor-alpha:

effects on

proliferation

of normal and trans-formed cells in vitro. Science

(Wash.

DC).

230:943-945.

11.

Kohase,

M.,D.

Henriksen-DeStefano,

L.

May,

J.

Vilcek,

and P.B.Sehgal. 1986. Induction of beta 2-interferon

by

tumornecrosis factor:ahomeostatic mechanism in the control of cell

proliferation.

Cell. 45:659-666.

12.

Sehgal,

P.B.,Z.Walther,and I. Tamm. 1987.

Rapid

enhance-mentof

,2-interferon/B-cell

differentiation factor BSF-2 gene expres-sion in humanfibroblastsby

diacylglycerols

and thecalcium iono-phoreA23187.Proc.Natl.Acad. Sci. USA. 84:3663-3667.

13.

Koeffler,

H.P.,J.

Gasson,

andA.Tobler.1988.

Transcriptional

and

post-transcriptional

modulation of

myeloid colony-stimulating

factor

expression by

tumornecrosis factor and other agents. Mol.

Cell.

Biol. 8:3432-3438.

14.Matsumoto, T.,T.

Hirano,

andT.Kishimoto. 1988. Establish-mentofaninterlukin-6

(IL-6)/B-cell

stimulating

factor

2-dependent

cellline and

preparation

of anti-IL-6 monoclonal antibodies. Eur. J. Immunol. 18:951-956.

15. Ponte, P.,P.

Gunning,

H.

Blau,

andL.Kedes. 1983.Human

actin genesare

single

copy of

alpha-skeletal

and

alpha-cardiac

actin but

multicopy

for beta- and

gamma-cytoskeletal

genes: 3' untranslated regionsare

isotype specific

butconserved in evolution.Mol.Cell.Biol. 3:1783-1791.

16.

Feinberg,

A.

_P.,

and B.

Vogelstein.

1983.Atechniquefor

ra-diolabelingDNArestriction endonucleasefragmentstohigh specific activity.Anal. Biochem. 132:6-13.

17.Clements, M.J. 1984. Purification of eukaryotic mRNA. In Transcription and Translation. B. D. Hames and S.J.Higgins, editors. IRLPress,Washington. 211-220.

18.Goldberg,D.A. 1980.Isolation and partial characterization of the Drosophila alcohol dehydrogenase gene. Proc. Nadl. Acad. Sci.

USA. 77:5794-5798.

19.Helfman,D.M., K. C. Barnes, J. M.Kinkade, W. M. Vogler, M.Shoji,and J. F.Kuo. 1983.Phospholipid-sensitive

Ca2l-dependent

proteinphosphorylation system in various types ofleukemic cells from humanpatientsand in humanleukemic cell lines HL-60 and K562 and its inhibition byalkyl-lysophospholipid. Cancer Res. 43:2955-2961.

20. Fujiki,H., M. Mori,M. Nakayasu,M.Terada, and S.

Sugi-mura. 1979.Apossiblenaturallyoccurringtumorpromotor, teleoci-din,3 from streptomyces. Biochem. Biophys. Res. Commun. 90:976-983.

21. Castagna, M.,Y.Takai,K.Kaibuchi,K.Sano,V. Kikkawa, andY.Nishizuka. 1982.Direct activation ofcalcium-activated, phos-pholipid-dependent protein kinase by tumor-promoting phorbol

esters.J. Biol. Chem. 257:7847-7851.

22. Sternweis, P.C., andA. C.Gilman. 1982. Aluminum:a

re-quirementfor activation of the regulatory component of adenylate cyclase by fluoride.Proc.Natl.Acad. Sci. USA. 79:4888-4891.

23.Solanki, V.,T. J.Slaga,M.Callaham, andE.Huberman. 1981. Down-regulation of specificbindingof[20-3H]phorbol

12,13-dibutyl-ateandphorbolester-induced differentiation of human promyelocytic leukemia cells.Proc.Natl.Acad. Sci. USA. 78:1722-1725.

24.Glyman,A.G., and S.J. D.Karlish. 1975. Thesodium pump.

Annu. Rev.Physiol.37:13-55.

25. Granger, G. A., G. E. Moore, J. White, P. Matzinger,J. S. Sundsma,S.Shupe,W. P.Kolb,J.Krammer,andP. R.Glade. 1970. Production oflymphotoxinandmigration inhibitingfactorby estab-lished humanlymphocyticcelllines.J.Immunol. 104:1476-1485.

26. Amino,N.,E.S.Linn,T.J.Pysher,R.Mier,G.E.Moore,and

L.J.DeGroot. 1974.Humanlymphotoxinobtained from established lymphoid cell lines: purification, characteristics and inhibition by anti-imunoglobulin.J.Immunol. 113:1334-1345.

27. Papermaster,B.W., 0.A.Holtermann,E.Klein,S.Parnett,D. Dobkin,R.Landico,and I.Djerassi.1981.Lymphokinepropertiesofa lymphoidculturedcellsupernatant fraction active inpromotingtumor regression.Clin. Immunol.

Immunopathol.

5:48-59.

28. Evans, C. H.,E.S.Rabin,and J.A.DiPaolo. 1977. The suscep-tibilityof Guineapigcellstothecolony-inhibitory

activity

of lympho-toxinduring

carcinogenesis.

Cancer Res.37:898-903.

29. Evans, C.H.,and J.A.Diaolo. 1981.

Lymphotoxin:

an anti-carcinogenic lymphokineasmeasuredbyinhibition of chemical

car-cinogenorultraviolet-induced transformation ofSyrianhamster cells. Int. J. Cancer. 27:45-49.

30.Gullberg, U.,M.Lantz,E.

Nilsson,

C.Peetre,G.

Adolf,

and I. Olsson. 1987.Characterization ofa

relationship

between the

T-lym-phocyte derived differentiation inducingfactor(DIF) and

lympho-toxin: acommon receptor system forDIF,lymphotoxinandtumor necrosisfactordownregulated byphorboldiesters. Eur. J. Haematol. 39:241-251.

31.Koeffler,H.P.,J.Gasson,J.Ranyard,L.Souza,M.

Shepard,

and R.Munker. 1987. Recombinant human

TNF-alpha

stimulates productionofgranulocytecolonystimulatingfactor.Blood.70:55-59. 32.Nishizuka,Y.1984. The role of

protein

kinase C in cell surface signaltransduction andtumor

_production.

Nature

(Lond.).

308:693-698.

33.Blackshean,P.M.,L. A.Witters,P.G.Girard,J. F.Luo,and S.N. Quamo. 1985. Growth factor-stimulated

protein

phosphoryla-tion in 3T3-L1 cells: evidence for

protein

kinase

C-dependent

and -independent pathways.J.Biol. Chem. 260:13304-13315.

(10)

immunoprecipi-tableprotein kinase C in GH3 cells treated with phorbol 12-myristate 13-acetate. J. Biol. Chem. 260:15194-15199.

35. Zhang, Y., J. X. Lin, and J. Vilcek. 1988. Synthesis of interleu-kin-6

(interferon-#2/B

cell stimulatory factor 2) in human fibroblasts triggered by an increase in intracellular cyclic AMP. J. Biol. Chem. 263:6177-6182.

36. Imamura, K.,M.L.Sherman,D.Spriggs, andD.Kufe. 1988. Effect of tumor necrosis factor on GTP binding and GTPase activity in HL-60 andL929 cells. J. Bid. Chem. 263:10247-10252.

37. Burch, R. M.,A.Luini, and J. Axelrod. 1986. PhospholipaseA2

and phospholipase Careactivatedby distinct GTP-binding proteins in responseto

alpha,-adrenergic

stimulation in FRTL5 thyroid cells. Proc. Natl. Acad. Sci. USA. 83:7201-7205.

38. Shaw, G., and R. Kamen. 1986. Aconserved AU sequence from the 3' untranslatedregionof GM-CSF mRNA mediates selective mRNAdegradation. Cell. 46:659-667.

39.Ikebuchi, K., G. G. Wong, S. C. Clark, J.N.Ihle,Y.Hirai, and

M.Ogawa. 1987. Interleukin-6 enhancement of interleukin-3-depen-dentproliferationofmultipotentialhematopoieticprogenitors. Proc. Natl. Acad. Sci. USA. 84:9035-9039.

40. Kishimoto, T. 1985. FactorsaffectingB-cell growth and differ-entiation.Annu.Rev.Immunol. 3:133-157.

41. Wood, D. D., E. J. Ihrie, C. A. Dinarello, andP.L. Cohen. 1983. Isolation ofaninterleukin-l-like factor fromhumanjoint effu-sions. Arthritis Rheum. 26:975-983.

42.Guerne, P. A., B. L. Zuraw, J. H. Vaughan, D. A. Carson, and

M.Lotz.1989.Synovium as a source of interleukin-6 in vitro. Contri-bution to local and systemic manifestations of arthritis. J. Clin. Invest. 83:585-592.

43. Hirano, T., T. Matsuda, M. Turner, N. Miyasaka, G. Buchan,

B.Tang, K. Sato, M. Shimizu, R. Maini, M. Feldmann, andT. Kishi-moto. 1988. Excessive production of interleukin-6/B cell stimulatory factor-2 in rheumatoid arthritis. Eur. J. Immunol. 18:1797-1801.

44. Dayer, J. M., B. Beutler, and A. Cerami. 1985. Cachectin/ tumornecrosis factor stimulates collagenase and prostaglandinE2 pro-duction by human synovial cells and dermal fibroblasts. J. Exp. Med. 162:2163-2168.