Factor-Dependent Macrophage

(1)

INFECTION IMMUNITY, Dec. 1980,p.797-804 0019-9567/80/12-0797/08$02.00/0

Role of Lipopolysaccharide in Regulating Colony-Stimulating Factor-Dependent Macrophage Proliferation In Vitro

ROBERT N. MOORE,* PATRICIA S.STEEG, DANIELA N. MANNEL,AND STEPHAN E. MERGENHAGEN

Laboratoryof Microbiology and Immunology, National Institute of DentalResearch, Bethesda, Maryland 20205

Bacterial lipopolysaccharides (LPS) enhancebothproduction ofcolony-stimulating factors (CSF)andproliferation of mononuclear phagocytes in vivo. The presentstudywasundertakentodeterminewhether the effects ofLPS onCSF- dependentmonopoiesisare due solelytoenhancedproduction of CSForalsoto directeffects of LPSontheresponding progenitorcell. Addition of LPStoCSF- stimulated macrophage populations had different effects, depending upon the concentration of CSFinthe cultures.In thepresence ofoptimaltosupraoptimal concentrations of CSF,LPSatdoses20.01

jig/ml

inhibitedmacrophage colony formation. Thisinhibitory activity was not due to cytotoxicity ofthe LPS and was notmediated through prostaglandinsynthesis. In the presenceofsuboptimal concentrations ofCSF, minute concentrations of LPS (l0-7 tug/ml) significantly enhanced macrophage colony formation. Both effects of LPS (inhibition and enhancement)appearedtobeproperties oflipidAsince neithereffectwasnoted with cells from LPS-resistant C3H/HeJ mice, whereas both effects could be neutralized by the addition of the antibiotic polymyxin

B,

which binds to the lipidAportion of LPS. These resultssuggestthat the effects ofLPSon mono- poiesisin vivomay not be duesolelytoits capacitytostimulate production of CSF. Rather, LPS maybeinvolved in stimulating monopoiesis both indirectly through stimulationofCSFproductionand by itseffects on theCSF-responsive progenitor cell.

Bacterial endotoxins and their lipopolysaccharide(LPS)moieties act on avariety of mam- malian cell types (reviewed in reference 3). Of these cell types, monocytes and macrophages areperhapsthe mostsensitive to the effectsof LPS. Teleologically, this is acceptable since thesecellsarethe mostlikelyto encountereither LPS or gram-negative bacteriain either endo- toxemicorinfected animals.

LPS has been shown to exert a variety of seeminglydirecteffectsonmacrophages. In ex- treme cases LPS can be cytotoxic for macrophages (14,33,37). In amajorityofexperimental systems, however, LPS is stimulatory rather than lethal. Thisstimulation has beendetected as enhancement of killing capacity (1, 10, 47) andassociatedproductionofoxygen metabolites (32), secretion of proteolytic enzymes (3, 46), andproductionof avarietyofsolublemediators (13, 23, 24, 26, 27, 43, 45),includingcolony-stim- ulatingfactor(CSF; 12).CSFisamyeloprolifer- ative agent (4, 5, 34) that has the capacity to stimulateclonalproliferationanddifferentiation ofmacrophageprecursors (38, 39). Both LPS (6, 8, 21,35) andinfection withgram-negativebacteria (2,40) enhance serumCSFlevels and stim-

ulate myelopoiesisin vivo. In aprevious report (28), weshowed thatLPS-induced production of CSF by macrophages is regulated by the con- comitant productionof Eprostaglandins. In the presentstudy,wehaveinvestigated the effectof LPS on CSF-induced cloning of macrophage precursorcells(39).Our findingsshow that LPS canexert different effects, depending upon the concentration ofCSFpresent inthe cultures. In thepresenceofoptimaltosupraoptimal concen- trationsofCSF, LPS inhibitedmacrophage col- onyformationin a noncytotoxic manner. In con- trast,macrophagecolony formation induced by suboptimalconcentrations ofCSFwas enhanced byaddition of minute amounts ofLPS.

MATERIALS AND METHODS Animals. MaleC3H/HeNmicewereobtainedfrom theNationalInstitutesofHealth Division of Research Resources, Bethesda, Md. C3H/HeJ mice of both sexes werepurchased from Jackson Laboratories,Bar Harbor,Maine. Theanimalswereusedat 6 to 8weeks ofage.

Reagents. LPS was extracted from Escherichia coliK-235bythephenol-watermethod ofMcIntireet al. (25). LPS solutionswereprepared fresh foreach experiment by suspending the LPS at 1 mg/ml in 797

(2)

798 MOORE ET AL.

Ca2+- and Mg2+-free Hanks balanced salt solution (HBSS; Microbiological Associates, Walkersville, Md).

Indomethacin and polymyxin B sulfate were pur- chased from Sigma Chemical Co. (St. Louis, Mo.).

Appropriate dilutions of reagents in HBSS were added to the cultures in avolume of0.1 ml. Conditioned medium (LCM) and column-fractionated CSF were prepared fromLcells byamodification of the method of Waheed andShadduck (42). Mouse L cells (clone 929;Microbiological Associates) weregrown in mini- mumessential medium(NIH Media Production Unit, Bethesda,Md.)supplemented with 5% fetal calfserum (MicrobiologicalAssociates) and antibiotics (50 U of penicillin and50

jig

ofstreptomycin perml). Confluent monolayers of cells weretrypsinized to remove the cells from their flasks. The cells were washed three timesinHBSS andsuspendedat 3 x 105 to 5 x 105 cells per ml inserum-free medium 199(NIH Media Production Unit)containing antibiotics.Atotal of50 mlof cellsuspensionwasthen addedto75-cm2 tissue culture flasks and incubated for7daysat370C ina5%

C02-in-air incubator. The supernatant medium was collected bycentrifugation and stored frozenat-20'C.

To prepare concentrated LCM, the medium was thawed and concentratedapproximately 200-fold in an ultrafiltration cell (Amicon Corp., Lexington, Mass.) with a PM-30 membrane.The concentratewasthen dialyzed for3daysagainstthree1,000-fold changes of distilled, deionizedwatercontaining100Uofpenicillin and 100

Itg

ofstreptomycinper ml. Thewater-soluble portionwasseparated bycentrifugationat10,000 x g anddialyzed against RPMI 1640mediumcontaining antibiotics. After sterilization by passage through0.22- Lmfilters(Millipore Corp.,Bedford, Mass.), the LCM wasstoredfrozen at-20'C.

To obtain a CSF-enriched preparation, the LCM (2ml)waschromatographedat4VCon aSephacrylS- 200(Pharmacia,Uppsala, Sweden) column (2.5 by95 cm) usingreverseflowat a rateofapproximately 12 ml/h. The buffer system used was 50 mM tris(hy- droxymethyl)aminomethane-hydrochloride (pH 7.5), containing0.1 MNaCI,and 3-ml fractionswerecol- lected. Themajority of the CSFwaseluted inarange correlating with marker proteins having molecular weights of 100,000 to 150,000. These fractions were pooled, concentratedto theoriginalvolume of LCM applied to the column, and dialyzed against RPMI 1640plus antibiotics. Theconcentrate wasthen filter sterilized and storedfrozen at-20'C.

Biological assays. CSFactivity wasdetermined insoftagar culture ofbone marrowcellsaspreviously described (28). Units ofCSF activity represent the calculatedcolony yieldin 1.0 ml ofsampleandwere determined fromthe linearportionof adilutionseries.

Cloning ofmacrophage precursor cells was meas- uredinliquid culture byamodificationof the method describedby Stewart and Lin(39). Thioglycolate-induced(2mlintraperitoneally,3days prior) peritoneal cellsfromC3H/HeNmice weresuspendedin enriched McCoy5Amedium(28)at 1x103cellsperml. A 2-ml amountof thiscellsuspensionwasthenplacedin 35- mmculturedishes(Falcon Division,Becton Dickinson

&Co.,Cockeysville, Md.) alongwith a sourceof CSF with or without otherreagents.Theplateswerethen incubatedat37°Cfor14daysin 5%C02-in-airatmos-

INFECT. IMMUN.

phere. Colonies (proliferative centers of 25 cells or more) were scored with a dissecting microscope at x30magnification. Colonies forming under these con- ditionsweregreaterthan 99% macrophages as judged bycellmorphology,differential staining, and phago- cytosisofantibody-coated sheep erythrocytes (41).

Direct visualassessment of proliferation and viability wasmadeby microscopic observation of cells cultured in eight-chambered slides (Lab-Tek no. 4838;

Lab-Tek Products, Div. MilesLaboratories Inc., Na- perville, Ill.). Peritoneal exudate cells (4 x 104) in 0.4 mlof McCoy medium wereadded to each well and allowedto adhere for2 h at37°C. After removal of nonadherent cells by washing, mediumcontaining dilutions of LCM orLPS orbothwasreturnedto the cultures. At timed intervals the chambers were in- spected withaninverted microscope atx200 magni- ficationfor cell numbers and forviabilityasmeasured by trypan blue exclusion (41). Based on randomly selectedsamples, both the primary and extended-term culturesweregreaterthan 98%macrophagesasjudged bymorphology, differential staining, andphagocytosis ofantibody-coated sheep erythrocytes.

Statistical analyses. Comparisons ofindividual means wereperformedwith the Student'sttest.

RESULTS

Inhibition

of macrophage colony formation by LPS. Addition ofLPStoLCM-stimulated adherentperitoneal exudate cells resulted in a dose-dependent inhibition ofmacrophage colony formation (Fig. 1).Thisinhibitoryeffect was mostdramaticatsupraoptimal dilutions of LCM andlessened,withlowerdoses ofLPS(1.0 to0.01

,ug/ml),

asthe dilution ofLCM increased.

LPSbyitselfdid not cause clone formation in these cultures.

Absence ofdetectable

cytotoxicity.

Cer- tainpreparations of LPS have beenshown to be cytotoxicformacrophagesat

relatively

highcon-

r

.)

0

MiGLPS/ML

0.01

LCMDILUTION-'

FIG. 1. Suppression by LPS ofmacrophageclone formationinduced by LCM. Adherent peritoneal exudate cells from 2 x 103 total exudate cells were exposedtodilutions of LCM in the presence or ab- senceof LPS. Colonies were counted 14 days later, and valuesrepresent the mean colony count of triplicate cultures.

(3)

EFFECTS OF LPS ON MONOPOIESIS IN VITRO 799 centrations (14).Although the LPS preparation

used in thisinvestigation was chosen because it was not toxicformacrophagesinvitro(<5% kill after overnight incubation of thioglycolate-in- ducedmacrophages with 50

jig

ofLPS per ml), it was considered necessary to determine whether the LPS was inhibiting macrophage colonyformationbykillingtherespondingcells.

Adirect microscopic analysis wasusedto ascer- tain this point. Exudate cells were cultured in eight-well chamber slides in the presence of LCMandLPS. Attimed intervalsthe medium was removed, and the cultures were observed for both viability, as measured by trypanblue exclusion,and for total cellnumbers. Asshown in Fig. 2A, the LPS alone did not exert any observable cytotoxic effect on any culture observed. The values shown in this figure repre- sentviable cellnumbers,which in allcaseswere

>98%. In these control cultures, to which no LCMwasadded,therewas anapproximatetwo- foldincrease incellnumbers perfieldduringthe 10-day observationperiod.This increasewas not significantly affectedby thepresence of up to 10

,Ig

of LPS per ml. In control cultures treated with either a 1:100 (20,000 U ofCSF) or 1:800 (2,500 U ofCSF) dilution of LCM,a significant proliferativeresponse wasobserved (Fig.2B and C),whichwasdramaticallyapparent by day10.

Inthose cultures treatedwith 1:100LCM (Fig.

2B),LPSexerted aninhibitory effectwhich was significantly different(P c 0.05) from the control atday 7. Inthesecultureslittleincrease in cell numbers occurred between days 7 and 10 al- thoughthere was nomeasurable increaseincell death.Inculturestreatedwith 1:800LCM (Fig.

2C),anobservableinhibition of proliferationwas notapparent untilday 10, and again the popu- lationremained greaterthan98%viable. Based ontheseresults,theinhibitory effectofLPS on

LCM-induced macrophage precursor division did not appear to be due to a direct cytotoxic effect of the LPS. TheLPS appeared instead to exert aCSF-dose-dependentcytostatic effect.

Evidence that the inhibitory effect is associated with lipid A.Macrophages from the C3H/HeJ mouse have been shown tobe either unresponsive or poorly responsive to a variety ofeffectselicitedby LPSand the lipid A portion ofLPS (11, 33, 41, 44). Inthis study, the exudate progenitor cells of this mouse strain also failed torespondto this inhibitory effect ofLPS(Fig.

3A) atconcentrations which are clearly inhibitory for the LPS-responsive C3H/HeN cells (Fig. 3B).

These results suggested that the inhibitory activity was mediated through lipid A. To test

300 A.C3H/HJ B.C3H/HN

P6LPS/M 0-00

260~~~~6j

220 A

FIG3.Coprtv0fetfLSo C-

100

60~ ^A

20

0 600 400 100 06400 1600 400 100 LCM DILUTIN

FIG. 3. Comparative effects of LPS on LCM-in- ducedcloningof macrophage progenitor cellsfrom C3H/HeJ (A) and C3H/HeN (B) mice. Valuesare colonycountsafter14daysofincubationofadherent cellsfrom2X 103 totalthioglycolate-inducedperitoneal exudate cells andrepresentmeansof triplicate cultures.The CSFactivity ofthe undiluted LCM used was2x 106U/ml.

2004 A OLCM

- -- -

a---,

B. 1:100 LCM

10 1 2 4 7

DZAYSINCULTURE

FIG. 2. Viablecellnumbersper X200microscopefieldattimed intervals after addition of LCM withor

without LPS. Values representmeansof fiverandomlyselectedfieldsandweregreaterthan 98% viableatall pointsasmeasuredbytrypanblue exclusion. Symbols forLPSdosesare asfollows(,ig/ml): 0(0); 0.01 (A);

1.0(0);10.0(-).LCMat1:100 and 1:800are20,000and2,500UofCSFactivityperml,respectively.

100

c 80

W 60

LLIq 40

w(-)I

C.1:800 LCM

20~ .I

1 2 10 1 2 4 10

in, ; -, .1 A .A . .A

VOLT 30,1980

(4)

800 MOORE ET AL.

this possibility, the LPS preparation was pre- treated with polymyxin

B,

which neutralizes LPS effects by forming a stable complex with thelipidAportion of the molecule (33). Addition ofthe antibiotic to the LPS blocked its inhibi- toryactivity for the sensitive mouse strain (Ta- ble1),providingfurther evidence that the inhib- itoryactivity of LPS is a property of the lipid A moiety.

Effect of indomethacin onthe

inhibitory

effect of LPS. Kurland and Moore

(19)

have recently shown that E

prostaglandins

inhibit CSF-induced cloning ofboth bonemarrowand peritoneal exudate

progenitor

cells. Since both LPS (18,43, 44)andCSF (38) stimulateE pros- taglandinproduction bymacrophages, thepos- sibility that these mediators

might

influence the cloningefficiencyofprogenitor cellsinthis

study

wasinvestigated.

Adherent peritoneal exudate cells were ex- posedtodifferent dilutions of LCM and LPS in thepresence orabsence of

indomethacin,

apo- tent inhibitor of the

cyclooxygenase pathway

of prostaglandin

synthesis.

The dose of indo- methacinused (1

Iug/ml)

hasbeen showntobe sufficient to block LPS-induced

synthesis

of

prostaglandins

by macrophages

(28).

As re- portedpreviously byHaddenetal.

(16),

addition of indomethacin to the cultures did have an enhancing effect on factor-induced

colony

formation(Fig. 4A).

Indomethacin, however,

failed to alter the inhibition of clonal

proliferation

effectedby the addition of LPS

(Fig. 4B).

Pros- taglandins,

therefore,

do not appearto

play

a major roleinthis

inhibitory

effect of LPS.

Enhanced

cloning

of

macrophage

progenitor cells with LPS and

suboptimal

concentrations of CSF. In several

experiments,

such as that shown in Fig. 2B, submicrogram

TABLE 1. Neutralizationofthesuppressiveeffectof LPS onmacrophagecolony formation

by polymyxin B

LPS PBb No. ofcolonies

Units ofCSF' ( g/ml)

(A9npl)

^perculturet

(~t/m1 (tg/1) SEW

2,500 0 0 179±5

2,500 1.0 0 82±3

2,500 0 10.0 164±15

2,500 1.0 10.0 192±31

0 1.0 0 0

0 0 10.0 0

aUnits ofSephacryl S-200 column-fractionated L cell-derivedcolony-stimulating activitypermilliliter.

b

Polymyxin

Bsulfate.

'Colonycount 14 daysafteradditionofCSFto 2 x 10' thioglycolate-elicited C3H/HeNperitoneal exudate cells. Valuesrepresent means oftriplicate cultures±standard error.

+INDO

2001

1001

I'-

0

UJn z 0 0

200

100

4

" I\

pjGLPS/ML

B 0.1--+ INDO-0-

10-*-+INDO-0-

/c

,43,

* / ,---- -0-~~~~~~~~~~~l

0 3200 600 800 400 200 100 -30 LCM DILUTION-1

FIG. 4. Effect of indomethacin(INDO,1pg/ml)on LCM-inducedcloningofmacrophage progenitorcells either in the absence (A) orpresence (B) ofLPS.

Valuesarecolonycountsafter14daysofincubation of adherent cellsfrom2x 103 thioglycolate-induced peritoneal exudate cells and representmeansof triplicate cultures. The CSF activity ofthe undiluted LCM usedwas2x106U/ml.

doses of LPS enhanced, ratherthan inhibited, macrophage cloning stimulated by suboptimal doses of LCM. Thissuggested that under certain circumstances LPS might enhance the cloning of macrophage progenitor cells. Since recent studies have shown that crude CSFpreparations might containhigh-molecular-weightenhancers forCSF-inducedcloninginvitro(M.Frolich,D.

W.

Golde,

and M. J.

Cline,

J.Supramol. Struct., suppl. 3. 1979, abstr. no 648, p.255), the LCM wasfractionated on a Sephacryl S-200column toobtaintheCSF-positive fractionsfreeofboth high- and low-molecular-weight contaminants.

Results obtained withcolumn-fractionated CSF aregiveninFig.5. In adose-dependentmanner, LPS could eithersuppress or enhance the clon- ingresponse ofC3H/HeNmacrophage progenitor cells, depending on the dose of CSF (Fig.

5A).Theinhibitory effectofLPSwasexpressed atoptimal andsupraoptimal concentrations of CSF, whereas the enhancing effect was most obvious in thesuboptimal tooptimal CSF con- centrationrange.Again,LPS hadnosignificant effect on CSF-inducedcloningofC3H/HeJmac- rophage progenitors (Fig. 5B).

Results shown in Fig. 6 establish two main points regardingtheenhancing effectofLPSon CSF-induced progenitorcell cloning. First, the progenitor cells are exquisitelysensitive to this INFECT. IMMUN.

300

^A

/0-- q

(5)

EFFECTS OF LPS ON 801

UNITSOF CSF/ML

FIG. 5. Comparative effects of LPS on CSF-in- ducedcloningof macrophage progenitor cellsfrom C3H/HeN(A)andC3H/HeJ(B)mice. TheCSFprep- aration usedwasthepooled fractions ofCSFactivity obtainedaftergel filtration ofLCMonSephacrylS- 200. Valuesarecolonycountsafter14daysofincu- bationofadherent cellsfrom2 x103totalthioglyco- late-inducedperitoneal exudate cells andrepresent meansoftriplicatecultures.

400o

-..,4 ..., 4

q-9 8-6 IC-7 C-6 m--5 C2-4 C

,,G _PS/YL

FIG. 6. Enhancedcloning of macrophageprecur- sorcellsexposedtoLPS and200 Uof column-frac- tionatedL cell CSFper ml. Values represented by closed circlesare meansoftriplicatecultures±one

standarderror,andvaluesrepresentedbyopencir- clesaresimilarmeansobtainedfromcultures incubated inthepresenceofpolymyxinBsulfate(PB)at 10 times the concentration ofeach LPSdose used.

Theshadedarearepresentsthemean±onestandard deviation ofthe number ofcoloniesformed in re- sponseonlyto200 UofCSFperml.

effectof LPS.In the presenceof200UofCSF,

aslittleas 10-7 ,ug of LPSpermlwassufficient tosignificantly enhance (P c 0.05) the cloning

response. Second, addition of polymyxin B to

theLPSatconcentrations10timesgreaterthan eachLPS dosewassufficienttototallyneutral- ize the enhancing effect of LPS at LPS doses _iO1-4,g/ml. At LPS doses >1i-4

Ig/ml,

polymyxin B was able to partially neutralize the enhancing effect. Polymyxin Batconcentrations

rangingfrom

10-8

to 10

,ug/ml

hadnoeffect on macrophage cloning stimulated by200U of CSF (datanotshown).The reversalby polymyxin B (Fig. 6 and Table 1), combined with the failure of C3H/HeJ cellstorespond toLPS (Fig. 5B), suggested thatthe enhancing effect ofLPS, as wellasthe inhibitory effect notedathigherCSF concentrations, wasdue to lipidA.

Kinetics of the LPS effects. The preceding

experiments

established that LPS can either enhance or suppress CSF-induced cloning of macrophage progenitor cells, depending upon the concentration ofCSF addedtothecultures.

Thefollowing experimentsweredesignedto determine the effect ofLPS when added before or after the additionofCSF.

Addition of LPS either 24 h

before

or after CSF had little effectoneithertheinhibitory or enhancing effect of LPSonthecloningresponse (Table 2). Addition of LPS delayedup to 6days after CSF also had no effect onthe inhibitory activity of LPS in thepresenceof either2,500 or 20,000 U ofCSF/ml (Table 3). In addition, de- layed addition of LPStoslide-chambercultures, although inhibiting proliferation, didnotresult in observable cell death (datanotshown), sup- porting the contentionthattheinhibitory effect ofLPS is notduetocytotoxicity. These results indicated thattheinhibitory effectofLPScould be exertedoncells already undergoingaprolif- erativeresponse torelatively high doses of CSF.

Enhanced cloning elicited by0.001

,ig

of LPS permlwasdiminished when the LPSwasadded afterCSF andwas

virtually

gonewhen the LPS wasadded6days later (Table3). TheLPSwas

apparently

influencing progenitor cellsinaman- nerthatallowed themtoproliferateinresponse to

suboptimal

doses of CSF. Upon

prolonged

culture these cellsgradually lost the capacityto respondtothiseffect ofLPS.

DISCUSSION

Both LPS and infection with gram-negative bacteria have profound effects on myelopoiesis in vivo. Sublethal doses of LPS increase both serumCSF levelsand numbers of mononuclear phagocytes andgranulocytes (6, 8, 21, 35).LPS hasalso been found to increase the numbers of myeloidstemcellsincellcycleintreatedanimals (21). Animals infected with gram-negative bacteria, such as Salmonella typhimurium, also show arapidincrease in serum CSF and a sub- sequent increase in numbers of mononuclear phagocytes (2, 40). Additionally, it has been shown that increases in serum CSF after X- irradiation are associated with the presence of gram-negativebacteria in the intestinal tract (7), whereas conventionalization of antibiotic-pre-

2^C

VOL. 30,1980

(6)

TABLE 2. Macrophage colonyformationbyexudate progenitor cells. exposed to LPS or CSF at0or 24 h Addition toculture at: No. ofcolonies perculture"at LPS dose(,Ag/ml)of:

Units ofCSF"

0h 24 h 0 1.0 0.01 0.0001

1,000 CSF+LPS -c 233 ±9 141 ±15 - -

CSF LPS 233 ±9 168±14 - -

LPS CSF 264±24 170 ±8 - -

200 CSF + LPS - 111 ±8 216±10 248±5 248±10

CSF LPS 111 ±8 222±9 270±12 254±16

LPS CSF 84±6 186± 14 228±6 244± 15

aUnits ofSephacryl S-200 column-fractionated L cell-derived colony-stimulating activity per milliliter.

'Colonycount 14days after addition of CSF to2 X103thioglycolate-elicited C3H/HeN peritoneal exudate cells. Values representmeansoftriplicatecultures±onestandard error.

c None.

TABLE 3. Effect of delayedadditionofLPSon macrophagecolonyformation byexudateprogenitor

cellsexposedtoLCMorSephacrylS-200- fractionated CSF

No. of colonies per cultureatday of Units of LPSdose LPSaddition':

CSFa (Ag/ml)

0 3 6

20,000' 0 201 ±11 -d -

1.0 93±18 89±3 90±24

2,500' 0 338±9 - -

1.0 170±2 178±5 171±7

200' 0 213±9 - -

0.001 362±20 307 37 232 35 a Units ofcolony-stimulating activityadded per culture.

'Colonycount 14daysafter addition of CSFto2x 10' thioglycolate-elicited C3H/HeN peritonealexudate cells.Val- ues represent meansoftriplicate cultures ± one standard error.

'Units ofLCM-CSFpermilliliter.

d-,0LPScontrol.

eUnits ofSephacrylS-200poolCSF.

treated canines resulted inanincrease inplasma CSF and a marked increase in bone marrow stemcells(22).Inallofthese cases, it isimplied thateitherLPSorother bacterialproductsare stimulating production of CSF which subse-

quently

enhances proliferation ofthe myeloid progenitor cells. Results reported in this com- munication, however, provided evidence that the effects of LPSon myelopoiesismaynotbe mediated solely through production of CSF.

Very small doses of LPS (10-7 Lg/ml)werefound tosignificantlyenhancecloningofexudatemac- rophageprogenitor cellsexposedtosuboptimal concentrations of CSF. This enhancing effect was most probably not due to production of additionalCSFin the cultures since it hasbeen our experience that adherent exudate cells do not produce detectable CSF when challenged with LPS. Thepossibilitythat exposuretoCSF might augment CSFproduction by LPS-stimulated exudate macrophages, however, has not

been eliminated. Thispossibility alsoseems un- likely since the number of clones forming in response to the combination of submicrogram dosesofLPS andsuboptimalconcentrations of CSF often exceededthat observed with optimal CSF doses (Fig. 5A). Therefore, the enhancing effectofLPSmaybe due to some other action ofLPS ratherthan to endogenous CSF production. Inthis context, the enhancement may be related to the observations of Dienstman and Defendi (9) that a significant portion of the exudatemacrophagepopulationcanbedirectly stimulated by LPS to enter the cell cycle.

Al-

though, as shown in our study, LPS alone is insufficient to permit a cloning response, it is possiblethat inthepresenceofsuboptimalcon- centrationsofCSF, LPSprovidesthe necessary stimulus for these cells to enterthe cell cycle while theCSF provides stimulationfor sustained proliferation.

LPS in thepresenceofoptimalorsupraopti- mal concentrations of CSF was found to suppress cloning of macrophage progenitor cells.

This inhibition is not the result of a cytotoxic effectof theLPS.Aswith theenhancing effect ofLPS, it is unclear whether this inhibitionis dueto adirect action ofLPS ontheprogenitor celloris mediatedbysome LPS-induced macrophage product. Three major factors which should be consideredas possible inhibitorsare E

prostaglandins

(19), tumornecrosis factor (36), and interferon (17). Addition ofindomethacin, which is sufficienttoinhibitprostaglandinpro- duction(28),failedtoaltertheinhibitoryeffect of LPS (Fig.2). This indicated thatprostaglan- dins, which can inhibit CSF-induced mono- poiesisin vitro(19),were notresponsiblefor the inhibitory effect of LPS in these experiments.

Recently, Shah et al. (36) presented evidence suggestingthattumornecrosis factorcaninhibit CSF-induced myelopoiesisin vitro. In arecent report from this laboratory (24), we presented evidence that CSF-cloned macrophages chal-

802

MOORE ET AL. INFECT. IMMUN.

(7)

lenged with LPS produce a cytotoxic factor with properties similar to serum-derived tumor necrosis factor. Addition of partially purified serum tumornecrosis factor (24) to the cultures, however, had noeffect on the CSF-induced cloning response (data notshown). This indicated that the inhibitoryeffectofLPS in these experiments wasprobably not due to the production of tumor necrosis factor in the cultures. Interferon has beenshown tohave anti-proliferativeactivities formammaliancells(15) and has been suggested to inhibit CSF-induced myelopoiesis in vitro (17). LPS is known to induceinterferon produc- tionby macrophages (23), and we have recently reported thatCSF-treated exudate macrophages produce enhanced levels of interferon when challengedwith LPS (29). In preliminary experiments, addition to the cultures of antiserum prepared against murine type 1 interferon (Na- tional Institute of Allergy and Infectious Dis- eases, Antiviral Substances Program) partially neutralizedtheinhibitoryeffect of LPS without affecting the normal cloning response to CSF.

Although more extensive investigation of the possible role of interferonin thisinhibitory effect ofLPSisrequired, it is clear that the effects of LPS,either inhibitory or enhancing, are related tothe dose ofCSF to which the cells are exposed.

With respect to interferon, it is important to note that a relationship does appear to exist between the effect of CSF as a macrophage- stimulatingagent and the resultsobtained in the present study. Exudate macrophages treated with at least 1,000 U of CSF display a signifi- cantlyenhancedresponsivenessto LPS as measured by interferon production (29). Doses of CSF lessthan 1,000 units yielded less than significant enhancement (29). Viewed in this context, it is possible that in thepresence of higher concentrations ofCSF the exudate population responds toLPSinan activatingfashion which precludes or inhibits proliferation, whereas LPS promotes a proliferative response in cultures exposed to suboptimal concentrations of CSF.

Regardless ofthe mechanism by which LPS either enhancesorinhibits CSF-induced cloning of macrophageprecursor cells, results obtained in thisstudy indicate that lipid A is probably the active portion of the LPS molecule in these effects.LPS neitherenhanced nor inhibited the cloning response of progenitor cells from the C3H/HeJmouse strainwhosemacrophages are genetically deficient in response to lipid A (11, 33, 41, 44). Also, both the enhancing and inhibitory effects of LPS on cells from genetically responsive mice could be eithertotallyor par- tiallyneutralized by polymyxin B, an antibiotic which bindstolipid A(31).

Although many

questions

remain to be re-

solvedregardingthe alterations of in vitromon- opoiesis by LPS, these results serve to reem- phasize the relationship between the mononuclear phagocyte system and LPS in both the infectedandhealthyhost.They clearlysuggest thatLPSisintimately involved inregulationof monopoiesis, both indirectly through stimula- tionofCSFproductionandbyits effectsonthe CSF-responsive

progenitor

cells.

ACKNOWLEDGMENTS

We thank JohnFarrar,TerryHoffeld, and Beda Stadler fortheir assistance in thepreparationofthismanuscript.We alsoacknowledgetheexcellentsecretarialhelp ofSybil Ceja.

LITERATURE CITED

1. Alexander, P., and R. Evans. 1971. Endotoxin and double stranded RNA rendermacrophagescytotoxic.

Nature(London)New Biol.232:76-78.

2. Bihme,D.H.,H. A.Schneider,and J. M. Lee. 1959.

Somephysiopathologicalparametersofnaturalresist- ance toinfectionin murinesalmonellosis. J.Exp. Med.

110:9-24.

3. Bradley,S. G.1979.Cellular and molecular mechanisms of action ofbacterial endotoxins.Annu.Rev.Microbiol.

33:67-94.

4. Bradley, T.R., and D. Metcalf. 1966. The growthof mousebonemarrowcellsinvitro.Aust. J.Exp.Biol.

Med.Sci.44:287-299.

5. Burgess, A. W.,D. Metcalf, and S. M. Watt. 1978.

Regulation ofhemopoietic cell differentiation andproliferation. J.Supramol. Struct.8:489-500.

6. Butler,R.C.,A. M.Abdelnoor,andA.Nowotny.1978.

Bonemarrowcolony-stimulatingfactorandtumorre- sistance-enhancing activity of postendotoxin mouse sera.Proc. Natl.Acad.Sci. U.S.A.75:2893-2896.

7. Chang, C.F., andM.Pollard.1973.Effects ofmicrobial floraonlevels ofcolonystimulating factor inserumof irradiatedCFW mice. Proc. Soc.Exp. Biol. Med.144:

177-180.

8. Chervenick, P.A. 1972. Effectofendotoxin andpost- endotoxin plasmaon in vitro granulopoiesis. J. Lab.

Clin.Med.79:1014-1020.

9. Dienstman, S. R., and V. Defend.1978.Necessaryand sufficient conditions for recruitment ofmacrophages intothecellcycle. Exp.Cell Res.115:191-199.

10. Doe, W.F.,and P. M.Henson.1978.Macrophagestim- ulation by bacterial lipopolysaccharides. I. Cytolytic effect on tumor targetcells.J.Exp.Med. 148:544-556.

11. Doe,W.F.,^{S. T.}Yang, D.C. Morrison,S.J.Betz, andP. M.Henson. 1978.Macrophage stimulationby bacteriallipopolysaccharides.II.Evidence for differen- tiationsignals delivered bylipidAandbyaproteinrich fraction oflipopolysaccharides. J. Exp. Med.148:557- 568.

12. Eaves,A.C.,and W. R. Bruce.1974.In vitroproduction ofcolony-stimulating activity. I. Exposure ofmouse peritonealcells to endotoxin. Cell Tissue Kinet. 7:19- 30.

13. Gery, I.,R.K.Gershon,and B. H.Waksman. 1972.

Potentiation of the T-lymphocytesresponse to mito- gens.I.Therespondingcell. J.Exp.Med. 136:128-142.

14.Glode,L.M., S. E.Mergenhagen,andD.L. Rosen- streich. 1977. Resistance ofmacrophagesfromC3H/

HeJ mice tothe in vitrocytotoxiceffectsofendotoxin.

J. Immunol. 119:162-166.

15. Gresser, I., J. De Maeyer-Guignard, M. G. Tovey, and E. De Maeyer. 1979. Electrophoretically pure mouseinterferonexertsmultiple biologiceffects. Proc.

VOL. 30,1980

(8)

804 MOORE ET AL.

Natl.Acad. Sci. U.S.A.76:5308-5312.

16. Hadden, J. W., J. R. Sadlik, and E. M. Hadden.1978.

The induction ofmacrophage proliferation in vitro by a lymphocyte-produced factor. J. Immunol. 121:231- 238.

17. Hull,E.V.,H.Schellekens,B.Lowenberg,and M. J.

deVries.1978. Influence of interferon preparations on theproliferative capacity of human and mouse bone marrowcells in vitro. Cancer Res. 38:911-914.

18.Kurland, J. I., and R. Bockman.1978.ProstaglandinE production by human blood monocytes and mouse per- itonealmacrophages. J.Exp. Med.7:952-957.

19.Kurland, J., and M. A. S. Moore.1977.Modulationof hemopoiesis byprostaglandins. Exp. Hematol. 5:357- 373.

20. Kurland, J. I., L.M.Pelus,P.Ralph,R. S.Bockman, and M. A. S. Moore.1979.Induction ofprostaglandin Esynthesisinnormal andneoplastic macrophages: role ofcolony-stimulating factor(s) distinct from effects on myeloid progenitor cell proliferation. Proc. Natl. Acad.

Sci. U.S.A. 76:2326-2330.

21. MacVittie,T.J.,and R. I. Walker. 1978.Endotoxin- induced alterations in canine granulopoiesis: colony- stimulating factor,colony-forming cells in culture and growth of cells in diffusion chambers.Exp. Hematol. 6:

613-618.

22. MacVittie, T. J., and R. L Walker.1978.Canine granulopoiesis: alterations inducedby suppression of gram- negative flora.Exp.Hematol. 6:639-647.

23. Maehara,N., andM.Ho. 1977.Cellularoriginofinter- feron inducedby bacterial lipopolysaccharide. Infect.

Immun.15:78-83.

24. Mannel, D.N.,R.N.Moore,and S. E.Mergenhagen.

1980.Endotoxin-inducedtumorcytotoxicfactor,p.141- 143. In D. Schlessinger (ed.), Microbiology-1980.

AmericanSociety forMicrobiology, Washington, D.C.

25. McIntire, F.C.,H. W.Sievert, G.H.Barlow,R. A.

Finley, and A. Y. Lee. 1967.Chemicalphysical and biological properties ofalipopolysaccharidefrom Esch- erichia coli K235.Biochemistry6:2363-2372.

26. Mizel,S.B.,J. J.Oppenheim,and D. L.Rosenstreich.

1978.Characterization oflymphocyte activatingfactor (LAF) produced byamacrophagecell lineP388D,. II.

Biochemical characterization of LAF inducedby acti- vated T cells andLPS. J. Immunol. 120:1504-1508.

27. Moore,R.N.,K. J.Goodrum,andL. J.Berry. 1976.

Mediationofanendotoxin effectby macrophages. RES J.Reticuloendothel. Soc.19:187-197.

28. Moore,R.N.,R.Urbaschek, L.M.Wahl,and S. E.

Mergenhagen. 1979.Prostaglandinregulationof colony-stimulating factor production by lipopolysaccharide-stimulatedmurineleukocytes. Infect. Immun. 26:

408-414.

29. Moore, R.N., S. N. Vogel,L. M. Wahl, and S. E.

Mergenhagen. 1980.Factorsinfluencing lipopolysaccharide-induced interferon production, p. 131-134. In D. Schlessinger (ed)., Microbiology-1980. American SocietyforMicrobiology, Washington, D.C.

30. Morland, B., and J. Morland.1978.Selective induction oflysosomal enzyme activitiesin mouseperitonealmac- rophages. RESJ.Reticuloendothel. Soc. 23:469-477.

31. Morrison,D.C.,andD. M.Jacobs.1976. Bindingof polymyxinB tothelipidAprotein ofbacteriallipopoly-

saccharide. Immunochemistry 133:813-818.

32. Pabst, M. J., and R. B. Johnston, Jr. 1980. Increased production ofsuperoxide anion by macrophages exposed in vitro to muramyl dipeptide or lipopolysaccharide. J.Exp. Med. 151:101-114.

33. Peavy, D. L., R. E. Baughn, and D. M. Musher. 1978.

Strain-dependent cytotoxic effects of endotoxin for mouse peritoneal macrophages. Infect. Immun.21:310- 319.

34. Pluznik, D. H., and L. Sachs. 1966. Theinduction of clones of normal mast cells by a substance from conditioned medium. Exp. Cell Res. 43:533-563.

35. Quesenberry,P., A. Morley, F. Stohman, Jr., K. Rich- ard, D. Howard, and M. Smith. 1972. Effect of endo- toxinongranulopoiesis and colony-stimulating factor.

N.Engl. J. Med. 286:227-232.

36. Shah, R. G., S. Green, and M. A. S. Moore. 1978.

Colony-stimulating and inhibiting activities in mouse serumafterCorynebacterium parvum-endotoxin treat- ment.RES J. Reticuloendothel. Soc.23:29-41.

37. Shands,J.W., Jr., D. L. Peavy, B. J. Gormus, and J.

McGraw. 1974. In vitro and in vivo effects of endotosin on mouseperitoneal cells. Infect. Immun. 9:106-112.

38. Stanely,E.R.,M.Cifone, P. M.Heard, and V. De- fendi.1976.Factorsregulating macrophage production andgrowth: identity of colony-stimulating factor and macrophage growth factor. J. Esp. Med.143:631-647.

39. Stewart, C. C., and H.iUn.1978.Macrophage growth factor and itsrelationship to colony stimulating factor.

RES J.Reticuloendothel. Soc. 23:269-285.

40.Trudgett, A., T. A.McNeil,and M.Killen.1973.Gran- ulocyte-macrophage precursor cell and colony-stimulating factor responses of mice infected withSalnonella typhimurium. Infect. Immun.8:450-455.

41.Vogel,S.N.,S. T.Marshall,and D.L. Rosenstreich.

1979.Analysis of the effects of lipopolysaccharide on macrophages:differential phagocytic responses of C3H/

HeN and C3H/HeJ macrophages in vitro. Infect. Im- mun.25:328-336.

42. Waheed, A., and R. K. Shadduck. 1979. Purification andproperties of L cell-derivedcolony-stimulating factor.J.Lab. Clin. Med.94:180-194.

43. Wahl,L M., C. E.Olsen,A.L.Sandberg,and S. E.

Mergenhagen. 1977. Prostaglandin regulation of macrophage collagenase production. Proc. Natl. Acad. Sci.

U.S.A.74:4955-4958.

44. Wahl, L M.,D.L.Rosenstreich, L. M. Glode, A. L.

Sandberg,and S. E.Mergenhagen. 1979.Defective prostaglandin synthesis by C3H/HeJ mouse macrophages stimulated with endotoxinpreparations. Infect.

Immun.23:8-13.

45.Wahl,L.M., S. M.Wahl, and J. B. McCarthy. 1980.

Adjuvantactivation ofmacrophage functions, p.491- 504.InE. R. Unanue andA.S. Rosenthal(ed.), Mac- rophage regulationofimmunity. Academic Press, Inc., New York.

46. Wahl, L. M., S.M.Wahl, S. E. Mergenhagen, and G.

R.Martin.1974.Collagenaseproduction by endotoxin activationmacrophages. Proc.Natl.Acad.Sci. U.S.A.

71:3598-3601.

47.Weinberg,J.B., H. A.Chapman, Jr., and J. B. Hibbs.

1978. Characterization of the effects ofendotoxinon macrophagetumorcellkilling. J. Immunol. 121:72-80.

INFECT. IMMUN.