role of tumor necrosis factor in macrophage activation and

[CANCER RESEARCH 48, 5427-5432, October 1, 1988]

Role of Tumor Necrosis Factor in Macrophage Activation andTumoricidal Activity1

Kenneth F. Mace,2 M. Jane Ehrke,3 Kazuyoshi Hori, Darbie L. Maccubbin, and Enrico Mihich

Grace Cancer Drug Center, Roswell Park Memorial Institute, New York Slate Department of Health, Buffalo, New York 14263

ABSTRACT

Tumor necrosis factor (TNF)-sensitive (LM) and -insensitive (P815)target cell lines were used to examine the role of TNF in both theactivation and lytic phases of macrophage-mediated lysis. LM cells werelysed spontaneously by thioglycolate-elicited macrophages in an 18-hassay (media or activating agents added with targets) or 36-h assay(macrophages cultured with media or activating agents for 18 h, washed,and targets added for a subsequent 18 h). In contrast, P815 cells werelysed only in the 36-h assay by macrophages exposed to appropriateactivation signals. Using antibody to murine TNF, it was shown that lysisof LM cells but not P815 cells was TNF mediated. The addition oflipopolysaccharide (LPS) to the 18-h assay resulted in augmented LM

killing. This was probably due to the fact that LPS stimulates macrophages to produce TNF. Conversely, when macrophages were pretreatedwith LPS for 18 h, washed, and assessed for lytic activity during thesubsequent 18 h, lysis of LM cells was reduced relative to the endogenouslevel.

Although macrophage lysis of P815 was not mediated by TNF, theaddition of TNF to macrophage activation cultures facilitated LPS triggering of cytolytic activity against P815. Similarly, the addition of TNFto the activation cultures partially prevented the LPS-induced reductionin macrophage-mediated LM cell lysis. Taken together, these data suggest that TNF may act as an autocrine signal during macrophage activation, in addition to being directly lytic to a select number of sensitivetarget cell lines.

INTRODUCTION

The in vitro activation of macrophages, obtained from avariety of tissues, to become cytolytic or cytostatic for tumorcells or microorganisms has been recently reviewed (1). Themolecular events which occur during the triggering and expression of macrophage-mediated cytotoxicity are receiving considerable attention (2). In several recent studies (3-8), it wasreported that TNF4 is elaborated by macrophages during the

activation process and that this is the primary mechanism bywhich activated macrophages kill tumor target cells. The demonstration that TNF was involved in macrophage-mediatedlysis was generally accomplished by blocking this reaction withantibody against TNF. However, these studies were all ratherlimited in scope since the target cell lines used were sensitiveto direct lysis by TNF. In studies on the growth of establishedcell lines in vitro, TNF was shown to either not effect growth,inhibit growth (as observed with L-cells), or enhance growth(9). Therefore, the proposal that tumoricidal macrophage activity is due to direct TNF lysis is limited to situations involvingTNF-sensitive tumor lines.

Received 2/23/88; revised 6/27/88; accepted 7/6/88.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

' Supported in part by Grants CA-15142, CA-24538, and CA-09072 awarded

by the National Cancer Institute, Department of Health and Human Services,and by a grant from Asahi Chemical Industry Co. Inc.

2This work was done in partial fulfillment of the Ph.D. degree requirements

in the Program of Pharmacology, Roswell Park Graduate Division, SUNY atBuffalo. Present address: Laboratory' of Molecular Immunoregulation, NCI-FCRF, 560/21-89A, Frederick, MD 21701-1013.

3To whom requests for reprints should be addressed.4The abbreviations used are: TNF, tumor necrosis factor; [3H]dThd, tritiated

deoxythymidine; LAK, lymphokine-activated killer (cells); Ik. lymphokine; LPS,lipopolysaccharide; NK, natural killer (cells); rH-TNF, recombinant human tumornecrosis factor, SN, supernatants: BCG, Bacillus Calmette-Guerin.

Numerous other reports have demonstrated divergent biological effects of TNF in a variety of experimental systems (10-12) and many investigators have reported that TNF can modulate various immune responses (3, 13-15). It was thereforereasonable to postulate that, under normal conditions, TNFfunctions as a regulatory cytokine (13). However, the relativephysiological importance of the various biological effects induced by TNF has yet to be clarified.

The present studies were designed to evaluate what functionsTNF may regulate during the activation and/or lytic phases oftumoricidal macrophage activity against both TNF-sensitiveand -insensitive target cells. Because of the numerous otherlytic factors produced by macrophages (e.g., reactive oxygen,neutral proteases, interleukin 1, complement, arginase, or factors inhibiting oxidative metabolism (16)), experimental conditions were chosen such that TNF production by macrophagescould be dissociated from macrophage killing. The results ofthis study, which have been reported previously in preliminaryform (17), confirm that TNF can serve as the lytic mediator inthe tumoricidal macrophage response against TNF-sensitivetargets; but, perhaps more importantly, it also serves as aregulatory molecule during the activation of tumoricidal macrophages. Furthermore, data are presented to support the postulate that tumor cell phenotype in part determines susceptibility to nonspecific effector cell lytic mechanisms.

MATERIALS AND METHODS

Mice. Specific pathogen-free female C57BL/6Cr mice were obtainedfrom Simmiscn Laboratories, Gilroy, CA, and were used for experiments at 8 to 12 weeks of age.

Culture Media. The medium used for all assays, macrophage incubations, and cell line maintenance was RPMI 1640 supplemented with0.1 mg/ml of gentamicin (Gibco Laboratories, Grand Island, NY), 25HIMAr-2-hydroxyethylpiperazine-A''-2-ethanesuIfonic acid buffer, and

10% fetal calf serum (HyClone Laboratories, Logan, UT). For thegeneration of LAK cells this medium was further supplemented with 1HIMglutamine, 0.1 IHMnonessential amino acids, 1 HIMsodium pyru-vate, and SOMM2-mercaptoethanoI (complete medium). Tissue culturemedia were endotoxin free (<O.OS ng/ml endotoxin), as determined byLimulus amebocyte lysate assay (M.A. Bioproducts, Walkersville, MD).

Reagents. rH-TNF was supplied by Asahi Chemical Industry, Tokyo,Japan, and the activity units of the preparations had been defined in anI. cell killing assay by the method of Yamazaki et al. (18). Rabbitantiserum against murine TNF-a was a gift provided through theextramural research program of Genentech, Inc., San Francisco, CA.For LAK generation, semipurified rat interleukin 2 was purchased fromCollaborative Research, Inc., Lexington, MA (Lot 861485). LPS (Esch-erichia coli 0111:B4) was purchased from Difco Laboratories, Detroit,MI; thioglycolate was from Becton Dickinson, Lincoln Park, NJ, andpolymyxin B was from Sigma Chemical Co., St. Louis, MO. [3H]dThd

(6.7 Ci/mol) was obtained from New England Nuclear, Boston, MA,and Na251CrO4(250-500 Â¿iCi/ngchromium) was obtained from Amer-

sham, Arlington Heights, IL.Cell Lines. L929 and its subline, LM, were obtained from American

Type Culture Collection, Rockville, MD. A TNF-resistant line (L929R)was developed by culturing L929 cells in the presence of rH-TNF (50units/ml) for 1 month. This line remained resistant after passage invitro, without further exposure to TNF. Half-confluent monolayers ofL-cells were harvested from culture flasks by gentle dislodgement by

5427

Research. on January 8, 2019. © 1988 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

TNF-MEDIATED MACROPHAGE ACTIVATION AND CYTOLYSIS

using a cell scraper (Costar, Cambridge, MA). Other cell lines usedwere P815 mastocytoma (syngeneic to DBA/2J mice) and YAC-1lymphoma (syngeneic to A/Sn mice). These cell lines grow in suspension cultures and were passaged three times per week.

L-Cell Assay of Supernatant Cytotoxic Activity. LM cells (104/0.1

ml) were cultured in 96-well plates for 48 h with supernatant (0.1 ml)from macrophages cultured for 18 h with or without LPS. Each wellwas pulsed with [3H]dThd (l ÃŸd/50 ^I/well) during the last 4 h of

incubation. Cells were harvested on glass filter paper by using an OttoHiller cell harvester. The radioactivity was measured by scintillationcounting with a Model LS7000 Beckman counter. Thymidine uptakeby LM cells was expressed as mean cpm Â±SD.

Isolation and Activation of Macrophages. Mice were given injectionsi.p. of 1 ml of 3% thioglycolate 5 days prior to sacrifice. PeritonealexÃºdate cells were harvested by lavage, using 7 ml of cold Hanks'balanced salt solution (Gibco). These cells (1.25 x 105/well) were added

to 96-well flat-bottomed plates (Corning Co., Corning, NY) in 0.1 ml.After a 2-h incubation at 37Â°Cin 5% CÃ›2, nonadherent cells were

removed by three washes and activation cultures were started immediately. The adherent population of cells was greater than 90% macrophages as determined by morphology and nonspecific esterase. Macrophages were cultured with or without agents (LK, LPS, TNF, poly-myxin B) for an 18-h activation period, washed, and "Cr-labeled targetcells (104/well) were added for a further 18-h incubation (36-h assay).

A variation of this assay used the coaddition of agents and radiolabeledtargets to adherent macrophages for 18 h (18-h assay).

NK and LAK Lytic Activities. Spleens were aseptically removed onday of sacrifice and single cell suspensions prepared by passage throughcoarse (50 mesh) and fine (200 mesh) stainless steel gauze. Viable cells,as determined by trypan blue exclusion, were counted on a hemacy-tometer. For cytotoxic NK assays, spleen cells (107/ml) were plated in

round bottomed microwell plates (Linbro, Flow Laboratories, McLean,VA) in a volume of 0.1 ml and 10" "Cr-labeled YAC-1 target cells were

added in 0.1 ml. The plates were incubated for 4 h, centrifuged, and0.1 ml of supernatant was removed from each well for determinationof radioactivity. For LAK generation, spleen cells (5 x 106/ml) werecombined with interleukin 2 (10 half-maximal units/well) and 1 MMindomethacin. LAK effector cells were harvested after 4 days of culture,counted, and tested for cytolytic activity (4 h) at 100:1 effectorrtargetratio (target cell number equaled 5 x 103/well). In each assay of NK

and LAK effectors at least 3 other effectontarget ratios were used andsimilar results were obtained (data not shown).

Evaluation of Specific "Cr Release. After 4 (NK, LAK) or 18 (mac

rophage) h of incubation of effector cells with target cells, the microwellplates were centrifuged at 300 x g for 5 min and the amount of 51Cr

released into 100 fi\ of cell free supernatant from each well was determined in a gamma counter. Percentage of specific 51Cr release was

calculated by the following formula:

% of specific 5lCr release

Experimental release - spontaneous releasetotal release - spontaneous release x 100

Experimental release is the radioactivity released in wells containingactivated effector cells and target cells; spontaneous release is theradioactivity released from target cells incubated in medium alone; totalrelease is the radioactivity from target cells lysed with 1% (v/v) TritonX-100. The spontaneous release was <10% in 4-h assays and <28% in18-h assays. Each test group was assayed in triplicate or quadruplicate

and all experiments were repeated at least three times.LK Preparation. Lymphokine-enriched culture supernatant of con-

canavalin A-stimulated mouse spleen cells was prepared as described(19). Briefly, spleen cells from untreated mice (7.5 x IO6 viable cells/

ml) were cultured for 24 h in complete medium containing 5 Mg/mlconcanavalin A (Pharmacia, Uppsala, Sweden). After the incubation,the supernatants were harvested, centrifuged, and treated with Sepha-dex G-10 (10 mg/ml) to remove residual concanavalin A from the LK.The LK was stored at 4Â°Cand its interferon activity was determined to

be 45 units/ml by a standard microtiter antiviral procedure (19).Tumor Necrosis Serum Preparation. As a primary stimulus for tumor

necrosis serum production, C57BL/6 mice were given injections i.v. of

2 mg of viable BCG (~6 x IO7 viable organisms) from Kyowa Hakko

Co., Tokyo, Japan. After 10 days, an i.v. injection of 50 MgLPS wasadministered to the BCG-treated mice. Blood was collected, by ventricle

puncture without anticoagulants, 2 h after LPS administration andserum was stored at â€”70Â°C.

Statistical Methods. All experiments were repeated at least threetimes with similar results obtained each time. Data from single experiments are shown expressed as mean values Â±SD. P values werecalculated by Student's t test.

RESULTS

rH-TNF-mediated Cytolysis of Murine Tumor Cell Lines.Four tumor cell lines were assayed for their sensitivity to rH-TNF in an 18-h5I Cr-release assay (Fig. 1). The L929 cell line,although sensitive to rH-TNF, was less sensitive than LM cellsand therefore the LM cell was used in the subsequent studies.L929R, an L929 subline resistant to rH-TNF, and P815, thestandard target cell line for assessing activated macrophagecytotoxicity, were resistant (in an 18-h assay) to rH-TNF at theconcentrations tested (5 x 10~4-50 units/ml). Using the rH-

TNF and the four targets under the same conditions in a parallel48-h postlabeling experiment (procedure as given in Fig. 3),similar results were obtained (data not shown).

Target Cell Cytolysis by Macrophages Pretreated with Different Activation Signals. Thioglycolate-elicited macrophages wereexposed to various factors for 18 h, washed, and recultured with5lCr-labeled tumor target cells for an additional 18 h. Macro

phages activated by LPS, LK plus LPS or LK plus TNF weresignificantly (P < 0.01) cytolytic toward TNF-insensitive P815cells (Fig. 2). Macrophages cultured with medium alone werecytolytic against TNF-sensitive LM cells and this spontaneouslysis was not affected by LK or LK plus TNF and was augmented by TNF alone. In contrast, preexposure of macrophagesto LPS or LK plus LPS significantly (P < 0.01) reducedsubsequent lytic activity against LM cells relative to that seenwith the untreated control. Under the conditions used, nocytolysis of L929R cells by macrophages was observed with anyof the test groups. A similar pattern of lysis was observed withresident macrophages; however, the absolute levels were lower(data not shown).

50 r

OV)

40 -

30

20

10

LM

U929

ADAO

O 5XIO"4 5XICT3 5XIO"2 5X10

rH - TNF (units

Fig. 1. Effect of rH-TNF on various murine tumor cell lines. LM, L929,P815, or L929R cells were labeled with "Cr. Tumor target cells (104/0.1 ml)were combined with rH-TNF (0.1 ml) in an 18-h "Cr-release assay. The datashown represent the average of triplicate wells. SD was <5% for all experimentalvalues. *, significant cytotoxicity (P < 0.001).

5428




bO403020P8ISLM-.Â»jÂ«i.1IÂ»â€¢T* L929Rfir1!

T T T TTCLK LPS TNF LK LK C LK LPS TNF L K LK C LKLPSTNFLK LK

+ + + + * *LPS TNF LPS THF Lps TNF

Table 2 Time dependence ofcytolysis by macrophages

Activation Signals Added In Culture

Fig. 2. Differential target cell susceptibility to macrophage pretreated withvarious activation signals. Macrophages were incubated for 18 h with media (<"i.

LK (1:25 dilution), LPS ( 1 Mg/ml), TNF (rH-TNF, 50 units/ml), or combinationsof these factors at the same concentrations, LK + LPS or LK + TNF. After thisactivation culture, the macrophage monolayers were washed prior to an I8-h"Cr-release assay by using labeled P815, LM. or L929R target cells. The resultsare averages of triplicate wells Â±SD. *, significantly increased cytotoxicity (/' <0.01) compared to control (C). **, significantly decreased cytotoxicity (P< 0.01)

compared to control (C).

Table 1 Abrogation of macrophage and tumor necrosis serum cytolysis by anti-murine-TNF antiserum

LM"

Lytic assay medium + antiserumElicited macrophages*

8h18hTNS

dilution''

1:2,000. 18h1:20,00030

Â±464Â±375.4

Â±2.953.3 Â±1.91.5

Â±0.5Â°4Â±2.5C4.0

Â±2.6C0 Â±1.3C

* Macrophages or TNS were mixed with medium or anti-murine-TNF antisera( 1:400 dilution) for 30 min prior to addition of "Cr-labeled LM cells for the lyticassay. The data presented represent the mean percentage of specific "Cr release

Â±SD.* Adherent macrophages from thioglycolate-elicited peritoneal exÃºdate cells

(1.25 x lO'/O.l ml) were incubated with radiolabeled LM target cells (IffVO.l

ml).' Significant reduction of lytic activity (/>< 0.01).'Final dilution of serum from BCG (1 mg, i.V., Day -10)-primed, LPS (25

ng, i.V., -2 h) induced C57BL mice.

Inhibition of TNF-dependent Tumor Cell Killing by Anti-TNFAntiserum. Antiserum raised in rabbits against murine TNFwas used to block spontaneous macrophage killing of LM cells(Table 1). TNF-sensitive LM target cells were added to unstim-ulated thioglycolate-elicited macrophages in the presence orabsence of antibody against murine TNF. In its absence, significant lytic activity was observed after 8 h and it doubled by 18h. The anti-TNF-antiserum abrogated this activity at both timepoints but had no effect on macrophage-mediated lysis of P815cells (data not shown). As a positive control for the specificityof this antiserum, it was also demonstrated that it abrogatedthe lytic activity (18 h) of tumor necrosis serum from BCG-primed, LPS-induced mice (Table 1).

Time Dependence of Cytolysis by Macrophages. In a 36-hassay, LPS stimulated macrophages to be lytic for P815 cellsbut inhibited spontaneous lysis of LM cells (Fig. 2). In order tounderstand these conflicting results, experiments were performed to determine the levels of macrophage-mediated cytolysis during and after exposure to LPS. Macrophages were: (a)incubated with labeled target cells Â±LPS (18-h assay); or (b)incubated Â±LPSfor 18 h (activation period), washed, and targetcells were added for 18 h (36-h assay). In both cases, target celllysis during an 18-h period was measured and is expressed aspercentage of specific 51Cr release (Table 2). Without LPSexposure, macrophages were cytolytic to TNF-sensitive LMcells, in both 18- and 36-h assays. Macrophage-mediated LMcytolysis was enhanced by LPS addition to the 18-h assay.

TargetcellLMP815LPS 18-hassay*48.7

Â±4.r-I- 63.8Â±1.6'0.6

Â±1.6+4.8 Â±2.236-h

assay*59.8

Â±1.520.9 Â±2.9'0

24 Â±3''

" Labeled target cells were added to macrophages Â±LPS (l Â«Â¿g/ml)and lyticactivity after 18 h was measured by "Cr release.

* Macrophages were preincubated for 18 h Â±LPS, then washed, prior to theaddition of labeled target cells. Lytic activity was assessed in an 18-h 91Cr-release

assay.c Data presented are the average percentage of specific "Cr release from

triplicate wells Â±SD.* Cytolysis in +LPS groups greater than -LPS group (P < 0.001 ).' Cytolysis in +LPS groups less than -LPS group (P < 0.001 ).

100

90tf)

80

2E 70OuOâ€” 60co

50

OO.

40

ocÃ‰30

10

0 10"' 10"

LPSFig. 3. Effect of macrophage culture supernatant on LM cells. Cell-free culture

SN from cultures of macrophage stimulated for 18h with the indicated concentrations of LPS were assayed for cytotoxicity against TNF-sensitive LM cells.LM cells (lO'/O.l ml) were cultured with SN (1:1, v/v) for 48 h; incorporation ofI'M|<lI lui d'tlR-i during the last 4 h was determined. Results are expressed as

average cpm of quadruplicate wells Â±SD. Incorporation of LM cultured withmedium alone was 96 x 10' Â±12.8 x IO3 cpm. *, significant cytotoxicity (P <

0.01).

Conversely, LM cytolysis mediated by macrophages in the 36-h assay was reduced when LPS, present during the activationperiod (the first 18 h), was removed prior to the addition oftarget cells. P815 target cells were not lysed in the 18-h assaywith or without LPS or in the 36-h assay without LPS activationof the macrophage. During the 36-h assay, significant macrophage-mediated cytolysis of the TNF-resistant P815 target cellsoccurred if LPS was present during the initial 18-h activationperiod.

Production of Growth-Inhibitory Soluble Factor by LPS-treated Murine Macrophages. Macrophages were cultured withor without LPS for 18 h. SN from these cultures were thenharvested and assayed for TNF-like cytotoxicity against LMcells (Fig. 3). LM cells were cultured for 48 h in the presenceof SN and were pulsed with [3H]dThd during the last 4 h.

5429




Reduced numbers of viable cells, as indicated by decreased[3H]dThd uptake, were observed with SN from LPS-treated

macrophage cultures but not from cultures without LPS addition. The elaboration of soluble cytotoxic factors increased withincreasing LPS concentrations.

Effect of Polymyxin B Addition to Macrophage ActivationCultures. In order to further establish the essential role of LPSin: (a) stimulating macrophage to lyse P81S; and (h) reducingspontaneous LM cell lysis, experiments using polymyxin B toneutralize the LPS activity were carried out (Table 3). Theaddition of polymyxin B (10 Mg/ml) blocked the LPS (1000 ng/ml) induction of P815 cell lytic activity and the LPS (10-1000ng/ml) induced down modulation of LM lysis.

Effect of TNF Addition to Macrophage Activation Cultureswith or without LPS. Previously, using similar experimentalconditions, TNF in combination with low concentrations of LK(19) or recombinant murine -y-interferon (20) was shown toactivate macrophages to kill TNF-resistant P815 cells. In orderto determine if TNF would have similar effects in combinationwith LPS, rH-TNF at various concentrations (0.05-5000 units/ml) was added to macrophage 18-h activation cultures with orwithout LPS (10 ng/ml). The cells were then washed, andlabeled target cells were added for the cytolytic assay. Significant activation of macrophage by TNF or LPS alone did notoccur at the concentrations tested (Fig. 4). When the two agentswere combined, however, a TNF concentration-dependent tu-

Table 3 Addition of polymyxin B lo macrophage activation culture

% ofcytotoxicityTumor

targetP815LMLPSÂ°(ng/ml)0

0.110

10000

0.1101000Control0Â±0.6

0Â±0.10Â±0.7

24.0 Â±3.4*52.2

Â±1.141.3 Â±9.610.7 Â±1.9*3.5 Â±0.6''+

polymyxinB0Â±0.2

2.3 Â±6.40Â±0.60Â±1.7*52.4

Â±6.447.0 Â±4.543.1 Â±2.2'3 1.5 Â±3.8'

" Adherent macrophages from thioglycolate-elicited peritoneal exÃºdate cells(1.25 x 105/0.1 ml) were incubated with or without LPS for 18 h (activation

culture) in the presence or absence of polymyxin B (10 Â¿tg/ml).The effector cellswere then washed and radiolabeled target cells (P815 and LM) were added. Thedata presented represent the average percentage of specific "Cr release after 18

h from triplicate wells Â±SD.* Cytolysis in +LPS group was greater than in -LPS group (P < 0.001).cCytolysis in +LPS + polymyxin B group was significantly less than that in

the +LPS control (-polymyxin B) group (P < 0.001).''Cytolysis in +LPS group was less than in -LPS group (P< 0.001).' Cytolysis in +LPS + polymyxin B groups was significantly greater than that

in the +LPS control (-polymyxin B) groups (P < 0.01).

0.05 05 5 50 500 5000

TNF (Units/mL)

Fig. 4. Effect of TNF on macrophage activation with LPS. Macrophages werecultured for 18 h Â±TNF (0.05-5000 units/ml) with (A) or without (â€¢)LPS (10ng/ml). Ina second 18-h incubation period, P8 15 target cells were added. Points,average percentage of specific "Cr release of triplicate wells; bars, SD. *, significant cytotoxicity (P < 0.01) of TNF + LPS versus TNF alone.

moricidal activation of macrophages was observed (P < 0.01).The data shown were obtained with thioglycolate-elicited macrophages but similar results were also observed with residentmacrophages.

Effect of Varying Time of TNF Addition to MacrophageActivation Cultures. It was apparent that even though P815 andLM macrophage-mediated lysis occurred through differentmechanisms, TNF had functional roles in activation and lyticevents, respectively. Experiments were performed, therefore, todetermine whether TNF exerted its effect on P815 and LMmacrophage-mediated cytolysis during both activation and lyticstages. As shown on Fig. 5, TNF (50 units/ml) addition at 0,8, or 18 h to the 24-h macrophage activation cultures containingLPS resulted in significant lysis (P < 0.05) of P815 targetsduring the subsequent 18-h assay culture (Fig. 5). However,when added at 24 h (after the LPS was washed out), TNF didnot effect P815 cell lysis.

As shown previously (Fig. 2), exposing macrophages to LPSduring the activation period resulted in a decrease (P < 0.01)in TNF-dependent LM lysis (Fig. 5). TNF alone augmented (P< 0.01) spontaneous macrophage-mediated LM lysis whenadded to these cultures at 0 and 8 h but not at 18 h. Furthermore, TNF addition (0, 8, or 18 h) reduced somewhat (P <0.05) the negative effect of LPS on spontaneous macrophagecytolysis of LM cells. As expected, when TNF was added withLM target cells to the 18-h lytic assay, enhanced cytolysis wasseen with both TNF (P < 0.01) and LPS/TNF (P < 0.01)treatment groups when compared to control and to LPS alonegroups, respectively. This increase in lysis of LM cells is likelyto be the combined result of lysis mediated by the exogenousTNF and macrophage-produced endogenous TNF.

Differential Target Cell Susceptibility to Cell-mediated Lysis.Experiments were performed to determine whether the L929Rcell line, which was resistant to TNF and to activated macrophage-mediated lysis, could be lysed by other nonspecific effector mechanisms such as LAK or NK (Table 4). In addition,these experiments examined the potential relationship betweentarget sensitivity to TNF and lysis by various effector cells.TNF-sensitive (LM) and TNF-resistant (P815 and L929R) celllines were used as targets as well as the YAC-1 cell line whichalso is resistant to rH-TNF (data not shown). Macrophageswithout LPS stimulation were only cytolytic to TNF-sensitive

P8I5

oo0)0)ce_p*Â»u8(/>aÂ«<Â°L.J ,11 .ft,.LM9294Â»80706050403020(0Â»

*_-â€¢11-

I10nftp.C

0 8 18 24

Time (h) of TNF Addition

Fig. 5. Effect of TNF addition at various times to macrophage activationculture. Macrophages were cultured for 24 h with medium (â€¢)or LPS (10 ng/ml, D). As indicated, certain cultures received TNF at various times, at 0, 8, or18 h during this 24-h activation period or together with targets at the end of theperiod (24 h) after all the cultures were washed. Lytic activity was then assessedin an 18-h "Cr-release assay. Columns, average percentage of specific "Cr-releasevalues of triplicate wells; bars, SD. *, significant cytotoxicity (P < 0.05) comparedto control. **, significantly decreased cytotoxicity (P < 0.01) of +LPS controlcompared to -LPS control. O, significant (P < 0.01) reversal of LPS-induceddecrease of cytotoxicity.

5430




Table 4 Differential target cell susceptibility to cell-mediated lysis

Effectorcells"Macrophage

-LPS+LPSLAK-IL-2'

+IL-2Target

cellsP8151.1

Â±1.2*25.1Â±1.3'1.9

Â±0.326.3 Â±!.</YAC-10Â±

1.313.3Â±1.4C0.8

Â±184.8Â±3.8rLM63.6

Â±5.414.5Â±6.5"1.2

Â±1.671.2Â±7.4/L929R0.6

Â±1.00Â±1.00Â±0.6

17.9 Â±\.</

NK 2.5 Â±3.6 27.9 Â±1.5 2.0 Â±0.9 2Â±3" The sensitivity of the different target cells (104/well) to lysis by three different

nonspecific effector cell populations. Macrophage cytotoxicity: macrophages wereincubated for 18 h Â±LPS (1 Mg/ml), washed, and MCr-labeled target cells were

added. LAK: spleen cells were cultured for 4 days with 1 y/Mindomethacin andÂ±IL2 (10 half-maximal units/ml), washed, counted, and added to labeled targetcells at an effectontarget ratio of 100:1. NK: fresh spleen cells were used aseffector cells to assay NK activity at the effectortarget ratio of 100:1.

* The data presented are the average percentage of specific 51Cr release Â±SDof triplicate wells after 4-h (LAK and NK) or IS h (macrophage) assays.

c Cytolysis in +LPS group greater than -LPS group (P < 0.001).d Cytolysis in -LPS group greater than +LPS group (P < 0.001).' IL-2, interleukin 2./Cytolysis in +IL-2 group greater than -IL-2 group (P < 0.001).

LM cells as shown previously (Table 2). With LPS-stimulatedmacrophages, significant lysis of P815, YAC-1, and LM tumortarget cell lines was observed. The L929R cell line was resistantto macrophage lytic effectors under all these experimentalconditions. When these same targets were assayed for susceptibility to LAK-mediated lysis, a different pattern of activitywas observed. The YAC-1 and LM target cells were lysed veryeffectively (>70% 51Cr release). Significant lysis although atlower levels occurred against P815 and even the macrophage-insensitive L929R cell line. Of the four lines tested for sensitivity to NK-mediated lysis, only YAC-1, the standard NK target,was significantly lysed.

DISCUSSION

Although tumor necrosis factor was first identified and hasbeen most extensively characterized in terms of its direct anti-tumor activity, the elaboration of TNF by macrophages canpotentially result in other physiological changes in vivo. One ofthe potential functions of TNF may be related to immuno-modulation and/or expression of host defense cytotoxic mechanisms against infectious disease and neoplasia.

The present studies were designed to evaluate possible function^) of TNF during the activation and lytic phases of thetumoricidal macrophage response as assessed against bothTNF-resistant and -sensitive target cell lines. Each of the 3target cell lines used in this study displayed a different pheno-type in terms of sensitivity to direct lysis by TNF and byactivated macrophages. Both the TNF-sensitive LM cells andthe inherently TNF-resistant P815 cells could be lysed byappropriately activated macrophage. However, the activationconditions differed for the two targets. Interestingly, the L929Rcell line which had an acquired resistance to TNF was not lysedin 18 h by any tested activated macrophage population. However, these cells were not simply resistant to lysis since theywere readily lysed by LAK cells. In fact, by testing target cellsusceptibility to lysis by NK and LAK in addition to macrophage, it was shown that the only apparent correlation betweenTNF sensitivity and susceptibility to lysis by nonspecific effector cells was with the LM cells.

It has been shown repeatedly that macrophage activation isstrictly dependent upon timing of exposure and concentrationof iÂ«vitro signals (1). The results of this study provide furtherevidence for this and show that the evaluation of tumoricidal

activation also depends upon the choice of target cell. Thus,the TNF-sensitive LM cells were lysed spontaneously in eitherthe 18- or 36-h assay. This spontaneous lysis was TNF mediatedsince it could be completely abrogated by anti-murine-TNFantiserum. Although the usual macrophage-activating signalswere not necessary during the activation stage of a 36-h assay,LM lysis was augmented when macrophages were exposed toTNF before the lytic period. This is probably due to the factthat TNF acts as an autocrine signal for macrophage to releasemore TNF (3) and/or to the binding of the exogenous TNF toreceptors on the macrophage plasma membrane (21). LPS haddiffering effects on macrophage-mediated LM lysis dependingon timing of exposure. LM lysis was augmented by LPS in the18-h assay but was significantly reduced in the 36-h assay (LPSpresent for 18 h, then washed out before lytic phase). Theaugmenting effect of LPS in the short term assay correlateswith the finding reported by others (4, 22) and confirmed inthis report that soluble factors with TNF activity can be elaborated by macrophages following LPS stimulation. The loss ofTNF-mediated LM lytic activity in the 36-h assay following

LPS exposure represents a down modulation of TNF expression. The removal of the positive stimulus supplied by LPS atthe end of the 18-h activation period may result in the downregulation of macrophage membrane-bound TNF (23), as wellas result in a decrease in soluble TNF production (22). Inaddition to a loss of the positive stimulus supplied by LPS, anegative regulatory mechanism such as that mediated by pros-taglandins may be involved. The addition of exogenous pros-taglandin E2 has been shown to suppress TNF production byLPS-stimulated macrophages (24). Furthermore, LPS is knownto be a potent signal for the elaboration of prostaglandin Â£2bymacrophages (25).

When P815 was used as a target cell, an entirely differentpicture emerged in terms of the conditions required for macrophage activation. P815 cells were only lysed in the 36-h assayand only by macrophages which had been exposed to appropriate activation signals (LPS alone, LK plus LPS, LK plus TNF).Under the conditions studied there was no evidence of LPS-induced down regulation of this P815 lysis. The tumoricidalactivation of macrophages by LK plus LPS or LPS alone hasbeen observed in numerous studies (1, 26). The activation ofmurine macrophages by LK plus TNF (19) or 7-interferon plusTNF (20) has been recently reported by this laboratory.

It was clear from these studies that when LPS is used as amacrophage-activating signal and then removed, it is possible

to demonstrate that macrophage tumoricidal activity through aTNF-mediated mechanism (LM) can be dissociated from thatexerted through a TNF-independent mechanism (P815). Thefact that LPS is essential for these effects to occur was furtherconfirmed by the observation that they could be blocked by thecoaddition of polymyxin B to the macrophage cultures.

Since it had been demonstrated that TNF, in cooperationwith other agents, acts as a macrophage activation signal (19,20), studies were initiated to determine if TNF would have asimilar positive effect on macrophage activation by LPS. Theaddition of TNF to macrophage activation cultures facilitatedLPS triggering of TNF-independent cytolytic activity (P815).These data further support the suggestion that TNF acts as anautocrine activation signal during macrophage activation. Previous studies, using polymyxin B, anti-TNF antiserum, andheat treatment had excluded the possibility that LPS contamination of the TNF preparation could be responsible for theobserved activation (19, 20). Talmadge et al. (27), using B16melanoma cells as targets, recently reported a similar TNF-induced activation of macrophages. Our study was different in

5431




that unlike the B16 melanoma, P815 is not sensitive to directlysis by TNF.

Any one or more of the various stages in a lytic mechanismsuch as recognition, binding, the production and/or secretionof lytic moieties by effector cells, or target cell sensitivity tothese lytic factors may be involved in determining a target cellresponse or lack of response to an effector. These phenomenaare reminiscent of the complex interactions which are knownto occur between target cell determinants and cytotoxic anti-cancer drugs (28). The recognition of these complex interactions together with the fact that tumor cells within a singletumor mass have been demonstrated to be heterogeneous interms of drug cytotoxicity (29) are fundamental to the conceptof multidrug combination therapy. If this concept can be appliedto immunotherapy, it can be suggested that a coordinatedactivation of multiple effector cell populations with differentlytic mechanisms will be required.

In summary, macrophages can be activated in vitro to atumoricidal state and the lytic profile of a population of activated macrophages is clearly dependent upon the activationsignals used, the timing of exposure to various activation signals, and the target cells used for assessing tumoricidal activity.This report indicates that both TNF-dependent and -independent lytic mechanisms occur within one, in v/fro-activated, macrophage population and that it is possible to dissociate the two.TNF-dependent macrophage lytic activity in this and previousstudies (3-8, 27) is limited to target cell lines directly sensitiveto the soluble form of TNF. It seems prudent at this time notto equate TNF cytotoxicity, which is very selective in nature,with either general macrophage-mediated cytotoxicity, which

has a much broader specificity, or with its therapeutic effects invivo (30-32) which may be mediated through its role as aregulatory cytokine. The fact that TNF-dependent and -independent macrophage lytic activities can be promoted by TNFshould be considered in both preclinical therapeutic studies andin clinical trials, especially if TNF is used in combination with7-interferon or other immunomodulators.

ACKNOWLEDGMENTS

We gratefully acknowledge the gifts of rabbit antiserum againstmurine TNF-a provided by Dr. M. Shepard, Extramural ResearchProgram, Genentech, Inc., and rH-TNF from Asahi Chemical IndustryCo., Tokyo, Japan. The authors wish to express their appreciation toKaren M. Schrader for secretarial assistance in the preparation of themanuscript.

REFERENCES

1. Adams, D. O., and Hamilton, T. A. The cell biology of macrophage activation. Annu. Rev. Immunol., 2: 283-318, 1984.

2. Drysdale, B. E., Yapundich, R. A., Shin, M. L., and Shin, H. S. Lipopoly-saccharide-mediated macrophage activation: the role of calcium in the generation of tumoricidal activity. J. Immunol., 138: 951-956, 1987.

3. Philip, R., and Epstein, L. B. Tumor necrosis factor as immunomodulatorand mediator of monocyte cytotoxicity induced by itself, 7-interferon andinterleukin-1. Nature (Lond.), 323: 86-89, 1986.

4. Austgulen, R., Espevik, T., Hammerstrom, J., and Nissen-Meyer, J. Role ofmonocyte cytotoxic factor in cytolysis of actinomycin D-treated WEHI 164cells mediated by freshly isolated human adherent mononuclear blood cells.Cancer Res., 46:4566-4570, 1986.

5. Nissen-Meyer, J., Austgulen, R., and Espevik, T. Comparison of recombinanttumor necrosis factor and the monocyte-derived cytotoxic factor involved inmonocyte-mediated cytotoxicity. Cancer Res., 47: 2251-2258, 1987.

6. Urban. J. L., Shepard, H. M., Rothstein, J. L., Sugarman, B. J., andSchreiber, H. Tumor necrosis factor: a potent effector molecule for tumor

cell killing by activated macrophages. Proc. Nati. Acad. Sci. USA, 83:5233-5237, 1986.

7. Feinman, R., Henriksen-DeStefano, D., Tsuiimoto, M., and Vilcek, J. Tumornecrosis factor is an important mediator of tumor cell killing by humanmonocytes. J. Immunol., Â¡38:635-640. 1987.

8. Ziegler-Heitbrock, H. W. L., Moller, A., Linke, R. P., Haas, J. G., Rieber,E. P., and Riethmuller, G. Tumor rÂ¿:Tosisfactor as effector molecule inmonocyte-mediated cytotoxicity. Cancer Res., 46: 5947-5952, 1986.

9. Sugarman, B. J., Aggarwal, B. B., Hass, P. E., Figari, I. S., Palladino, M.A., Jr., and Shepard, H. M. Recombinant human tumor necrosis factor-a:effects on proliferation of normal and transformed cells in vitro. Science(Wash. DC), 230: 943-945, 1985.

10. Old, L. J. Tumor necrosis factor (TNF). Science (Wash. DC), 230:630-632,1985.

11. Mannel. D. N. Biological aspects of tumor necrosis factor. Immunobiology,172: 283-290, 1986.

12. Beutler, B., and Cerami, A. Cachectin and tumor necrosis factor as two sidesof the same biological coin. Nature (Lond.), 520: 584-588, 1986.

13. Ehrke, M., Mace, K., Maccubbin, D., and Mihich, E. Regulatory role ofhuman recombinant tumor necrosis factor on murine, control and Adria-mycin-modified, host defense functions. In: S. Eckhardt (ed.). Abstract ofLectures, Symposia and Free Communications of the XlVth InternationalCancer Congress, Hungary, Vol. 3, p. 916, Basel: Karger, 1986.14. KasliÂ¡Â«a.H., Wright, S. (.'...and Bonavida, B. Regulation of B cell maturation

and differentiation. I. Suppression of pokeweed mitogen-induced B celldifferentiation by tumor necrosis factor (TNF). J. Immunol., 138: 1383-1390, 1987.

15. Klebanoff, S. J., Vadas, M. A., HarÃan,J. M., Sparks, L. H., Gamble, J. R.,Agosti, J. M., and Waltersdorph, A. M. Stimulation of neutrophils by tumornecrosis factor. J. Immunol., 136:4220-4225, 1986.

16. Nathan, C. F. Secretory products of macrophages. J. Clin. Invest., 79: 319-326, 1987.

17. Mace, K., Ehrke, M. J., and Mihich, E. Dissociation of target cell sensitivityto tumor necrosis factor and macrophage-mediated cytolysis. Fed. Proc., 46:1388, 1987.

18. Yamazaki, S., Onishi, E., Enami, K., Natori, K., Kohase, M., Sakamoto, H.,Tanouchi, M., and Hayashi, H. Proposal of standardized methods andreference for assaying recombinant human tumor necrosis factor. Jpn. J.Med. Sci. Biol., 39:105-118, 1986.

19. Hori, K., Ehrke, M. J., Mace, K., Maccubbin, D., Doyle, M. J Otsuka, Y.,and Mihich, E. Effect of tumor necrosis factor on the induction of macrophage tumoricidal activity. Cancer Res., 47: 2792-2798, 1987.

20. Hori, K., Ehrke, M. J., Mace, K., and Mihich, E. Effect of tumor necrosisfactor on tumoricidal activation of macrophages: synergism between tumornecrosis factor and f-interferon. Cancer Res., 47: 5868-5874, 1987.

21. Bakouche, O., Ichinose, Y., Heicappell, R., Fidler, I. J., and Lachman, L. B.Plasma membrane-associated tumor necrosis factor: a non-integral membrane protein possibly bound to its own receptor. J. Immunol., 140: 1142-1147, 1988.

22. Gifford, G. E., and Lohmann-Matthes, M. L. Requirement for the continualpresence of lipopolysaccharide for production of tumor necrosis factor bythioglycollate-induced peritoneal murine macrophages. Int. J. Cancer, .M.135-137, 1986.

23. Decker, T., Lohmann-Matthes, M-L., and Gifford, G. E. Cell-associatedtumor necrosis factor (TNF) as a killing mechanism of activated cytotoxicmacrophages. J. Immunol., 138:957-962, 1987.

24. Kunkel, S. L., Wiggins, R. C., Chensue, S. W., and Larrick, J. Regulation ofmacrophage tumor necrosis factor by prostaglandin E2. Biochem. Biophys.Res. Commun., 13 7:404-410, 1986.

25. Taflet, S. M., and Russell, S. W. Macrophage-mediated tumor cell killing:regulation of cytolytic activity by prostaglandin E. J. Immunol., 126: 424-427, 1981.

26. Russell, S. Stimulate the phagocytes. Immunol. Today, 8: 347-349, 1986.27. Talmadge, J. E., Tribble, H. R., Pennington, R. W., Phillips, H., and

Wiltrout, R. H. Immunomodulatory and imtnunotherapeutic properties ofrecombinant tumor necrosis factor in mice. Cancer Res., 47: 2563-2570,1987.

28. Rustum, Y. M., Grindey, G. B., Hakala, M. T., and Mihich, E. Multifactorialcellular determinants of the action of antimetabolites. Adv. Enzyme Regul.,14:281-295, 1976.

29. Slocum, H. K., Heppner, G. H., and Rustum, Y. M. Cellular heterogeneityof human tumors: implications for understanding and treating cancer. In: E.Mihich (ed.), Biological Responses in Cancer, Vol. 4, pp. 183-227. NewYork: Plenum Press, 1985.

30. Chiara, P., Boriiseli!, D., Nencioni, L., Ghezzi, P., and Tagliabue, A. Enhancement of in vivo immune response by tumor necrosis factor. J. Immunol.,139: 3676-3679, 1987.

31. Chun, M., and Hoffmann, M. K. Combination immunotherapy of cancer ina mouse model: synergism between tumor necrosis factor and other defensesystems. Cancer Res., 47:115-118, 1987.

32. Manda, T., Shimomara, K., Mukumoto, S., Kobayashi, K., Mizota, T., Hirn!.O., Matsumoto, S., Oku, T., Nishigaki, F., Mori, J., and Kikuchi, H.Recombinant human tumor necrosis factor-a: evidence of an indirect modeof antitumor activity. Cancer Res., 47: 3707-3711, 1987.

5432



1988;48:5427-5432. Cancer Res Kenneth F. Mace, M. Jane Ehrke, Kazuyoshi Hori, et al. Tumoricidal ActivityRole of Tumor Necrosis Factor in Macrophage Activation and

Updated version

http://cancerres.aacrjournals.org/content/48/19/5427

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/48/19/5427To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



role of tumor necrosis factor in macrophage activation and

Documents