reduced antilisterial activity of tnf-deficient bone marrow-derived macrophages is due to impaired...

0014-2980/99/1010-3089$17.50+.50/0© WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999

Reduced antilisterial activity of TNF-deficient bonemarrow-derived macrophages is due to impairedsuperoxide production

Matthias Müller1, Roland Althaus2, Dieter Fröhlich3, Karl Frei4 and Hans-PietroEugster5

1 Novartis Pharma Research, Basel, Switzerland2 Biological Research Institute, Matzingen, Switzerland3 University Hospital, Department of Anaesthesia, Regensburg, Germany4 Department of Neurosurgery, University Hospital Zürich, Zürich, Switzerland5 Department of Internal Medicine, Section of Clinical Immunology, University Hospital Zürich,

Zürich, Switzerland

Mice deficient for TNF ligand and receptor type 1 have demonstrated the importance of TNFin the host defense against Listeria monocytogenes. To investigate the particular deficiencyof macrophages derived from TNF/lymphotoxin (LT)- § −/− mice in antilisterial growth control,bone marrow-derived macrophages (BMDM) were used for in vitro infection experiments.After the combined treatment with IFN- + and lipopolysaccharide (LPS), production of NO bywild-type (wt) and TNF/LT- § −/− BMDM was induced to comparable levels, but only wt BMDMcontrolled L. monocytogenes growth efficiently. Nevertheless, inhibition of NO productionled to a remarkable loss of antilisterial activity. This suggests that presence of NO is neces-sary but not sufficient for L. monocytogenes killing and that elimination of L. monocytogenesrequires additional effector molecules. The LPS-inducible superoxide production of TNF/LT-§−/− BMDM was impaired. Accordingly both scavenging of superoxide and peroxynitrite led

to reduced L. monocytogenes killing by wt BMDM. In addition, peroxynitrite was able to killL. monocytogenes in vitro. Together these findings suggest that the defective host defenseof TNF/LT- § -deficient mice against L. monocytogenes partially stems from reduced superox-ide production of macrophages due to the absence of TNF and imply a function for peroxyni-trite, the reaction product of NO and superoxide, in the intracellular killing of L. monocyto-genes.

Key words: Macrophage / Host defense / Nitric oxide / Superoxide / Peroxynitrite

Received 3/5/99Revised 25/6/99Accepted 2/7/99

[I 19618]

R. Althaus and M. Müller have contributed equally to thiswork.

Abbreviations: BMDM: Bone marrow-derived macro-phages L-NMMA: NG monomethyl-L-arginine MnTMPyP:Mn(II/III)-tetrakis-(1-methyl-4-pyridyl)-porphyrin iNOS: In-ducible nitric oxide synthetase ICSBP: Interferon consen-sus binding protein IRF2: Interferon regulatory factor 2wt: wild-type MESF: Molecule equivalents of soluble fluo-rochrome LT: Lymphotoxin ROI: Reactive oxygen inter-mediates

1 Introduction

In concert with T cells, neutrophils and NK cells, macro-phages are critically involved in resistance against infec-tion with Listeria monocytogenes [1–5]. At the molecularlevel, TNF, IL-1, IL-6, IFN- + and NO have been shown tobe crucially involved in the murine defense against L.

monocytogenes [6–9]. All four cytokines can enhanceresistance to experimental infection of mice with L.monocytogenes when administered before infection [9].In addition, in vivo administration of antibodies againstthe different cytokines resulted in increased bacterialgrowth and eventual death of mice from listeriosis[10–12]. Mice deficient for IFN- + R [13], TNFR1 [14], TNFand lymphotoxin (LT)- § [15], interferon consensus bind-ing protein (ICSBP), interferon regulatory factor 2 (IRF2)[16] and NADPH oxidase [17] show gradually reducedsusceptibility towards L. monocytogenes infection andtherefore demonstrate the importance of IFN- + TNF andreactive oxygen intermediates (ROI) for efficient hostdefense.

At the macrophage level, TNF has been shown to pro-mote in vitro killing of L. monocytogenes [8]. IFN- + isrequired for induction of inducible NO synthetase (iNOS)and NO generation in macrophages [18]. Inhibition of NO

Eur. J. Immunol. 1999. 29: 3089–3097 Effector molecules in murine antilisterial activity 3089

Figure 1. TNF and NO dependence of L. monocytogeneskilling by BMDM. (A) Bacterial killing was determined in wtand TNF/LT- § -deficient BMDM after pretreatment of BMDMwith 200 U/ml IFN- + , 10 ng/ml TNF, 10 ng/ml LPS as well asthe NO synthetase inhibitor L-NMMA (1 mM) and a TNF-neutralizing serum ( § -TNF, 10 % v/v) for 18 h in differentcombinations. (B) Nitrite production was determined fromthe corresponding assay supernatants. Detection limit wasfound to be 2 ? M (dashed line) and data are given asmean ± SD of triplicate assays. All experiments were per-formed at least three times and a representative data set isgiven.

synthesis blocks antilisterial activity in vitro [19–21] andin vivo [22] and iNOS-deficient mice are susceptible tohigh inocula of L. monocytogenes [23], defining NO as acritical antilisterial effector molecule. In contradiction tothis finding is that mice deficient for IRF1, which are notable to produce NO [16, 24], are not susceptible to L.monocytogenes. The combined requirement for ROI andNO for a competent host defense against L. monocyto-genes, however, was demonstrated in mice deficient foriNOS and NADPH oxidase [25]. These findings put inperspective the importance of NO in the host defenseagainst L. monocytogenes and favors a predominant roleof ROI in antilisterial host defense. Since TNF as well asIFN- + are prototype inducers of superoxide [8, 26–29],we hypothesized that TNF-mediated superoxide produc-tion might be critical for an effective antilisterial hostresponse.

Here we describe that lack of TNF/LT- § leads todecreased antilisterial activity of bone marrow-derivedmacrophages (BMDM) and that this deficiency is sug-gested to be due to reduced superoxide production andas a consequence might impair the NO-dependent andNO-independent killing pathway of macrophages.

2 Results

2.1 Endogenously produced TNF and NO arecrucial effector molecules in the intracellulargrowth control of L. monocytogenes byBMDM

Since mice deficient for TNF/LT- § were unable to controllow titer infection with L. monocytogenes [15], we inves-tigated the role of endogenously produced TNF in anti-listerial growth control of BMDM in vitro. TNF/LT- § −/− andwild-type (wt) BMDM were prestimulated with IFN- + andLPS for 18 h and assayed for antilisterial growth control.wt BMDM prestimulated with IFN- + and LPS reducedbacterial survival by 80 % whereas TNF/LT- § −/− BMDMonly showed a reduction of bacterial survival of 20 %compared to unstimulated BMDM from both genotypes(Fig. 1A). In the same assays using LPS and IFN- + presti-mulation, NO production was comparable between wtand TNF/LT- § −/− BMDM. Dependence of listericidal activ-ity upon TNF was further demonstrated by an abrogationof growth control by neutralization of endogenous TNFwith anti-TNF antisera in wt BMDM prestimulated withIFN- + and LPS (Fig. 1A). These results clearly demon-strate the importance of endogenous TNF for the antilis-terial growth control of BMDM.

Replacing LPS by TNF during prestimulation of BMDMfrom wt and TNF/LT- § −/− mice resulted in a 90 % reduc-

tion of intracellular L. monocytogenes growth, demon-strating that exogenous TNF can restore the impairedantilisterial growth control to wt levels (Fig. 1A). Individu-ally, prestimulation of BMDM with TNF or IFN- + showedonly a limited but nevertheless clear effect in growth con-trol despite undetectable NO production (Fig. 1A, B).Inhibition of the iNOS by NG monomethyl-L-arginine (L-NMMA) reduced antilisterial activity by 40 % and 65 % inwt and TNF/LT- § −/− BMDM, respectively. This supportsan important function of NO in antilisterial growth controlof BMDM (Fig. 1B).

2.2 Impaired superoxide anion production ofBMDM from TNF/LT- § –/– mice

The production of hydrogen peroxide was determinedbased on rhodamine-123-related fluorescence and isexpressed as molecule equivalents of soluble fluoro-

3090 M. Müller et al. Eur. J. Immunol. 1999. 29: 3089–3097

Table 1. Respiratory burst activity of wt and TNF/LT- § −/− BMDM after different stimulia)

wt TNF/LT- § −/−

Prestimulation Stimulus Absolutefluorescence

¿ fluorescence (%) Absolutefluorescence

¿ fluorescence (%)

None None 3.9 ± 0.1 −84.3 5.6 ± 0.8 −74.5

None L. monocytogenes 24.7 ± 0.8 0 21.9 ± 2.0 0

IFN- + PMA 75.0 ± 7.0 203.6 66.2 ± 9.7 202.2

IFN- + L. monocytogenes 21.8 ± 1.9 −11.8 21.7 ± 3.2 −0.9

IFN- + /LPS L. monocytogenes 28.1 ± 1.3 13.7 20.6 ± 4.5 −5.9

IFN- + /TNF L. monocytogenes 36.3 ± 6.9 46.9 37.8 ± 3.1 72.6

a) After prestimulation with cytokines and LPS for 18 h, cells were loaded with the fluorogenic substrates for 10 min, then theoxidative burst was elicited with 5 × 106 CFU L. monocytogenes or 100 nM PMA. The results of the cellular fluorescence aregiven in 103 MESF. ¿ fluorescence (%) gives the inducible oxidative burst relative to the values generated by listeria alone (2ndrow). Data are given as mean ± SD of three independent experiments.

Figure 2. Superoxide and peroxynitrite scavenging impairs killing of L. monocytogenes by BMDM. Inhibition of bacterial killingof wt and TNF/LT- § −/− BMDM pretreated for 18 h with 200 U/ml IFN- + and 10 ng/ml LPS or 10 ng/ml TNF by the superoxidescavenger MnTMPyP (A) or the peroxynitrite scavenger uric acid (100 ? M) (B). Data are given as mean ± SD of triplicate assays.

chrome (MESF). During the oxidative response of phago-cytes, superoxide anions are generated by a membrane-bound NADPH oxidase. This oxidase transforms molec-ular oxygen through a single electron transfer into super-oxide anions. Without prestimulation, BMDM of eithergenotype showed equal basal levels of fluorescenceelicited by L. monocytogenes (Table 1). PMA induced astrong oxidative burst in IFN- + -prestimulated BMDM ofeither genotype, demonstrating that the enzymesrequired for oxidative burst activity are also present inTNF/LT- § −/−. In contrast to the L. monocytogenes-elicited superoxide production of IFN- + /LPS-pretreatedwt BMDM, the corresponding superoxide production ofTNF/LT- § −/− BMDM did not differ from baseline. Treat-ment with exogenous TNF together with IFN- + normal-ized the oxidative response of TNF/LT- § −/− BMDM to wtvalues (Table 1).

2.3 Superoxide and peroxynitrite scavengingreduce antilisterial growth control

Peroxynitrite, the reaction product of NO and superoxideanion, has been suggested to be an ultimate effectormolecule in killing of Escherichia coli [30]. By using thecell-permeable superoxide dismutase mimic Mn(II/III)-tetrakis-(1-methyl-4-pyridyl)-porphyrin (MnTMPyP) as asuperoxide anion scavenging agent [31] and uric acid asa peroxynitrite scavenger [32], the superoxide anioneffector arm as well as the suggested final effector mole-cule were tested for their importance in listerial growthcontrol. BMDM of wt and TNF-deficient mice were pre-stimulated with IFN- + and either LPS or TNF, and treatedwith either MnTMPyP or uric acid. Treatment of wtBMDM with MnTMPyP led to a dose-dependent reduc-tion in antilisterial activity independent of the prestimula-tion with IFN- + and LPS or TNF (Fig. 2A). In the case of


Figure 3. In vitro bactericidal activity of peroxynitrite. Expo-nentially growing L. monocytogenes were incubatedtogether with various concentrations of peroxynitrite for60 min at 37 °C. Peroxynitrite kills L. monocytogenes in adose-dependent manner. Survival data are given asmean ± SD of triplicate assays.

TNF/LT- § −/− BMDM, the inhibitory effect of MnTMPyPwas observed only with the highest dose if BMDM wereprestimulated with IFN- + and LPS. However, if TNF/LT- § –/– BMDM were prestimulated with IFN- + and TNF, aninhibition became apparent at the lower MnTMPyP con-centration. In an analogous experiment, uric acidreduced the antilisterial growth control in wt BMDM two-fold but showed no effects in TNF/LT- § −/− BMDM whenstimulated with IFN- + and LPS (Fig. 2B). Only marginaluric acid-mediated inhibition of growth control wasobserved when BMDM were stimulated with IFN- + andTNF. These data suggest that both superoxide anion aswell as peroxynitrite play a pivotal role in macrophage-mediated killing of L. monocytogenes.

2.4 Peroxynitrite efficiently kills L. monocyto-genes in vitro

To test if peroxynitrite itself could be an ultimate effectormolecule of murine macrophages, the peroxynitrite-mediated killing was investigated in vitro. Peroxynitriteadded to L. monocytogenes cultures during their expo-nential growth phase was able to efficiently kill the bacte-ria in a dose-dependent manner (Fig. 3). No killing wasobserved with vehicle alone (data not shown).

3 Discussion

Studies on knockout mice revealed the importance ofTNF signaling via TNFR1 in the host defense againstL. monocytogenes [14, 15, 17, 33].

On the basis of our in vitro killing assays we were able toshow that the impaired L. monocytogenes killing of TNF/LT- § −/− BMDM can be restored by exogenous TNF andthat neutralization of IFN- + /LPS-induced TNF in wtBMDM led to strongly decreased L. monocytogenes kill-ing. Thus, the positive para- and autocrine mediatorfunction of TNF in antilisterial activity of BMDM has beenshown and supports earlier in vivo studies in murine lis-teriosis [34].

Earlier experiments with a macrophage cell line [21],iNOS-deficient mice [23] and knockout mice deficient inIFN- + signaling which have impaired iNOS induction [13,24] suggested that NO plays an important role in the hostdefense against L. monocytogenes. The use of L-NMMAblocked measurable NO production completely andincreased the L. monocytogenes survival from 15 % toabout 55 % in wt BMDM, a survival which is also seen inwt BMDM prestimulated with IFN- + alone. The sole pro-duction of NO, however, did not result in antilisterialactivity since TNF-deficient BMDM generating similar

amounts of NO as wt BMDM after IFN- + and LPS presti-mulation (Fig. 1B) showed only a limited (30 %) killing ofL. monocytogenes (Fig. 1A). In addition, neutralization ofTNF by antibodies led to a 60 % reduction of killing byLPS- and IFN- + -stimulated BMDM in the presence ofhigh NO levels (Fig. 1B). A similar unresponsiveness ofNO generation to TNF neutralization has been describedby others [19].

The fact that inhibition of antilisterial growth control by L-NMMA was more pronounced with BMDM from TNF/LT-§−/− mice than from wt mice suggests an additional anti-

listerial, TNF-dependent but NO-independent killingmechanism. Possible candidate effector moleculesmight be ROI such as superoxide, hydrogen peroxideand hydroxyl radicals [35]. Indeed, it has been shownthat inhibition of NO synthesis by L-NMMA increasedH2O2 production via dismutation of superoxide in murinemacrophages [27] and might be responsible for theresidual L. monocytogenes killing of L-NMMA-treated wtBMDM. Such an NO-independent killing pathway for L.monocytogenes has been described with a mouse mac-rophage precursor hybrid cell line [36]. In addition, Fehret al. [16] have shown that peritoneal macrophages fromICSBP-deficient mice cannot kill L. monocytogenesdespite their ability to produce NO like wt mice, favoringthe importance of ROI over the reactive nitrogen interme-diate pathway in L. monocytogenes killing.

For effective antilisterial activity, IFN- + and a secondstimulus such as TNF or LPS was essential. This interre-


Figure 4. TNF- and IFN- + -controlled NO-dependent andindependent killing pathways in BMDM. Schematic presen-tation of the TNF- and IFN- + -controlled oxidative pathwaysshowing membrane-associated NADPH oxidase, xanthineoxidase and cytosolic NO synthetase associated to thephago-lysosomal compartment generating the precursormolecule of peroxynitrite in the immediate vicinity of phago-cytosed L. monocytogenes. L-NMMA and uric acid inhibitthe NO-dependent pathway of L. monocytogenes killing,whereas MnTMPyP inhibits both the NO-dependent andNO-independent pathway. The dependence of superoxideand NO generation from the combined action of TNF andIFN- + has been shown by analysis of L. monocytogenes-elicited oxidative burst and bacterial killing assays, respec-tively.

lationship between TNF and IFN- + as outlined in Fig. 4might in part be based on the fact that TNF is able toenhance IFN- + R expression on monocytes [37] and IFN- +might act via its IL-10 down-regulatory properties onmonocytes [38] and thereby allow higher TNF expression.Peritoneal macrophages or macrophage-like cell linesperform antilisterial activity upon IFN- + -stimulation alone[16, 19, 21]. This is not the case with wt BMDM whichonly showed a limited killing of 35 %. This suggests thatperitoneal macrophages represent cells in an advanceddifferentiation state towards a host defense phenotypewhere the basal requirements for an efficient effectorstate might be reduced. Recently, the requirement forautocrine TNF action for IFN- + -induced NO generation ofmurine peritoneal macrophages has been demonstratedby the complete inhibition of IFN- + -induced NO produc-tion by a TNF-neutralizing antiserum [39]. This supportsthe above assumption and suggests that TNF precondi-tions peritoneal macrophage preparations.

The consequences of TNF/LT- § −/− deficiency in BMDMbecame apparent by the lack of LPS-induced superox-ide production in TNF/LT- § −/− BMDM (Table 1). The IFN- +dependence of the oxidative response of murine perito-neal macrophages has been shown recently by Fehr etal. [16], however without investigating the function ofTNF. Endres et al. [17] investigated the N-formyl-Met-Leu-Phe (fMLP)-elicited oxidative burst of granulocytesfrom TNFR1-deficient and control mice. These authorsdid not demonstrate a defect in the oxidative burst ofTNFR1−/− granulocytes. This is possibly because fMLPbypasses the TNF-controlled pathway of superoxideproduction. A similar, normal oxidative response wasseen in our TNF/LT- § −/− BMDM if stimulated with PMAand IFN- + (Table 1).

In the context of the impaired oxidative burst of TNF/LT-§−/− BMDM, the consequences of blocking the oxidative

pathways on L. monocytogenes killing of BMDM wereinvestigated. Scavenging superoxide with MnTMPyP wasvery effective in blocking antilisterial growth control medi-ated by IFN- + and LPS or TNF (Fig. 2A). In contrast, theperoxynitrite scavenger uric acid only partially blockedantilisterial growth control in IFN- + - and LPS-stimulatedwt BMDM (Fig. 2B), the residual killing being 50 % whichwas equal to the residual killing observed in wt BMDMprestimulated with IFN- + and TNF and treated with theiNOS inhibitor L-NMMA. This identical effect of either1 mM L-NMMA, which blocks NO generation completelyand 100 ? M uric acid which scavenges peroxynitrite sug-gests that about 50 % of the L. monocytogenes killingmight stem from the NO-dependent effector pathway.

Obvious no effect of uric acid was observed on antiliste-rial growth control of TNF/LT- § −/− BMDM if prestimulated

with IFN- + and LPS, suggesting that no peroxynitrite isformed under these conditions. However, if BMDM ofboth genotypes were prestimulated with IFN- + and TNF,only a marginal inhibition of antilisterial growth control byuric acid became apparent. These observations suggestthat the prestimulation of BMDM with IFN- + and LPS orTNF led to different stages of activation. This was thecase since the oxidative burst elicited in BMDM of eithergenotype by L. monocytogenes was fourfold higher ifBMDM were prestimulated with IFN- + and TNF than ifBMDM were stimulated with IFN- + and LPS (Table 1).Therefore, it is conceivable that uric acid was able toscavenge the amount of peroxynitrite produced in wtBMDM prestimulated with IFN- + and LPS, but did notreach an intracellular concentration to scavenge theamount of peroxynitrite produced by BMDM prestimu-lated with IFN- + and TNF.

Scavenging superoxide by MnTMPyP theoreticallyblocks the two diverging pathways of L. monocytogeneskilling outlined in Fig. 4, the NO-dependent peroxynitrite


pathway and an NO-independent superoxide pathway.Based on our findings we can conclude that peroxynitritein part contributes to the killing of intracellular L. mono-cytogenes. This conclusion is supported by the efficientin vitro killing of L. monocytogenes by peroxynitrite(Fig. 3) and by the data of Shiloh et al. [25] which showedthat macrophage-mediated listeria killing is mainly con-trolled by the NADPH oxidase pathway and that micedevoid of functional NO synthase and NADPH oxidaseare highly susceptible to infection by L. monocytogenes.

In conclusion these results suggest that the generationof NO and superoxide in murine BMDM requires bothsignals from IFN- + and TNF and demonstrate that botheffector pathways are closely connected. NO and super-oxide can generate peroxynitrite which is an ultimateeffector molecule in the killing of L. monocytogenes.Inhibiting the oxidative pathway at the level of superox-ide is most efficient because the NO-dependent and NO-independent pathway as depicted in Fig. 4 are blockedand finally contributes to complete impairment of L.monocytogenes killing in wt BMDM. Inhibition at thelevel of iNOS or peroxynitrite targets only the NO-dependent killing pathway and results only in a partialblock of the L. monocytogenes killing capacity. Consid-ering that macrophages represent only a part of the anti-listerial host defense we can nevertheless conclude thatTNF/LT- § deficiency leads to reduced superoxide pro-duction. The same reduction in effector moleculeexpression might also impair the early antilisterial func-tion of neutrophils [2] and generally weaken the reactivenitrogen intermediate/ROI-mediated host response inTNF/LT- § -deficient mice.

4 Materials and methods

4.1 Mice

TNF/LT- § double-deficient mice [15] were bred in-house andwere of a mixed genetic background (C57BL/6 × 129).C57BL/6 × 129 mice were used as controls.

4.2 Reagents

LPS from E. coli (serotype 0111:B5) was purchased fromSigma (St. Louis, MO) and resuspended in pyrogen-freesterile PBS. L-NMMA salt was purchased from Calbiochem-Novabiochem AG (Laeufelfingen, Switzerland). L. monocy-togenes strain EGD, and a polyclonal sheep antiserumagainst recombinant murine TNF [40], were kindly providedby Dr. R. M. Zinkernagel (Institute of Experimental Immunol-ogy, University Hospital, Zürich, Switzerland). Recombinantmurine TNF was from Genzyme (Stehelin, Basel, Switzer-land) and recombinant mouse IFN- + (3.6 × 106 U/mg) was a

generous gift from Dr. L. Ozmen (Hoffman-La Roche AG,Basel, Switzerland). 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) was purchased fromSigma. Dihydrorhodamine 123 (DHR) and carboxy-seminaphthorhodafluor-1-acetoxymethylester (SNARF1/AM) were from Molecular Probes (Eugene, OR); DHR wasdissolved in dimethyl formamide at a concentration of10 ? M; SNARF1/AM at 1 ? M. Aliquots were stored at−20 °C. MnTMPyP and freshly synthesized peroxynitrite in0.3 M sodium hydroxide were from Alexis (ALEXIS Corpora-tion, Laeufeldingen, Switzerland). Peroxynitrite was storedat −80 °C and used for bacterial killing assays within 2weeks after shipment.

4.3 Bactericidal assay

To obtain BMDM, bone marrow cells were flushed fromfemurs of 8-week-old mice and cultivated in conditionedmedium in bacterial petri dishes for 8 days [41]. BMDM weredetached in cold PBS at 4 °C and resuspended at a densityof 106 cells/ml in conditioned medium. Aliquots of 5 × 104

cells were transferred to flat-bottom 96-well plates and pre-treated for 18 h with different cytokines (IFN- + , 200 U/ml;TNF, 10 ng/ml) and LPS (10 ng/ml) alone or in combinationwith TNF antibodies (200 000 neutralizing U/ml; 10 % v/v) orthe iNOS inhibitor L-NMMA (1 mM) [42]. The superoxide andperoxynitrite scavenger MnTMPyP [31] and uric acid [32]were added to the BMDM cultures at the same time as thebacteria and were present until termination of the experi-ments. L. monocytogenes from a fresh overnight culturewere added in 100 ? l DMEM in a ratio of 3 CFU/cell. Cul-tures were incubated for the indicated time period at 37 °C in5 % CO2. After carefully washing the cultures with pre-warmed medium four times, titers of intracellular bacteriawere assayed immediately. Briefly, intracellular bacteriawere released by lysing macrophages with 10 ? l 5 % sapo-nine (v/v) (Sigma) for 2 min at room temperature. After addi-tion of LB medium (200 ? l/well), plates were incubated for3 h at 37 °C in 5 % CO2. The resulting number of bacteriawas determined on the basis of MTT conversion asdescribed by Leenen et al. [36]. Filter-sterilized MTT (10 ? l;5 mg/ml) was added to the culture and incubated for 30 min.Formazan formation was measured at 620 nm using a Lab-systems Multiskan® Biochromatic ELISA plate reader (Lab-systems, Shrewsbury, MA). To validate the MTT-basedassay, the numbers of bacteria were determined in parallelby plating diluted cultures on LB agar plates. Both methodsgave consistent and reproducible results. Peroxynitrite-mediated killing was performed in microtiter plates. Peroxy-nitrite in 0.3 M sodium hydroxide was added to emptymicrotiter wells, then 5 × 105 exponentially growing L.monocytogenes were added in a volume of 200 ? l LB brothand the reaction was allowed to proceed for 60 min at 37 °C.The relative survival of bacteria was determined by MTTconversion as described above.


4.4 Determination of nitrite concentration

NO was measured as nitrite using the Griess reagent [43].Culture supernatants (50 ? l) were mixed with 100 ? l 1 % sul-fanilamide and 100 ? l 0.1 % N-(1-naphthyl)-ethylenediaminedihydrochloride and 2.5 % orthophosphoric acid (all Sigma).Absorbance was measured at 570 nm after 20 min in anELISA plate reader (Labsystems, Shrewsbury, MA). Nitriteconcentration was determined with reference to a standardcurve using concentrations from 1 to 250 ? M sodium nitrite(Sigma) in culture medium. Reliable sensitivity was at 2 ? M.All data represent mean values ± SD from three individualexperiments.

4.5 Oxidative response

The non-fluorescent dihydrorhodamine 123 is oxidized intra-cellularly in a peroxidase-dependent reaction to green fluo-rescent rhodamine 123, emitting green light (520–540 nm)upon excitation at 488 nm, which is measurable by flowcytometry. In polypropylene tubes (diameter 10 mm) 5 × 105

BMDM were stimulated with different cytokines and LPS asmentioned above. After 18 h BMDM were washed once withPBS. The BMDM were loaded with the fluorogenic sub-strates DHR and SNARF1/AM for 10 min at 37 °C. Then L.monocytogenes at a ratio of 10:1 or 100 nM PMA as a posi-tive control were added to initiate the oxidative response ofthe BMDM. After 1-h incubation the reaction was stoppedon ice. Dead cells were counterstained with propidiumiodide at a final concentration of 30 ? M. The probes werestored on ice in the dark and measured within 1 h. The finalconcentrations were 1 ? M for DHR 123 and 0.1 ? M forSNARF1/AM. For analysis we used a FACScan flow cyto-meter (Becton Dickinson, San Jose, CA) with argon ion laserexcitation 488 nm, measuring 10 000 cells of each sample.Data were acquired and processed using LYSIS-II software.Based on calibration with dye-beads (Quantum 26, Flowcy-tometry Standards Europe, Leiden, The Netherlands), theresults of the cellular fluorescence were expressed in MESF.For this the dye-beads of known fluorochrome content wereassessed after each set of measurements. The resultingdata were used to calculate a calibration curve based on lin-ear regression. Based on this curve the arbitrary flow cyto-meter units were converted to MESF. These MESF unitswere used for the absolute quantification of cellular fluores-cence, allowing inter-assay and inter-laboratory comparisonof data. Leukocyte esterase activity was assessed usingSNARF1/AM. SNARF1/AM is cleaved in viable leukocytesby esterases to SNARF1, a pH-sensitive dye. The intracellu-lar presence of SNARF1 fluorescence allows the discrimina-tion of vital macrophages from cellular debris. Beside thelacking esterase activity, propidium iodide fluorescence(above 600 nm) was used as an additional criterion to detectdead cells. The percentage of dead cells was less than 1 %in all assays.

Acknowledgements: We thank A. Fontana for continuoussupport as well as helpful discussions and critical readingof the manuscript. This work received financial supportfrom the Swiss National Science Foundation grant 32-33966.92.

5 References

1 Kaufmann, S. H., CD8+ T lymphocytes in intracellularmicrobial infections. Immunol. Today 1988. 9: 168–174.

2 Rogers, H. W. and Unanue, E. R., Neutrophils areinvolved in acute, nonspecific resistance to Listeriamonocytogenes in mice. Infect. Immun. 1993. 61:5090–5096.

3 Dunn, P. L. and North, R. J., Early gamma interferonproduction by natural killer cells is important in defenseagainst murine listeriosis. Infect. Immun. 1991. 59:2892–2900.

4 Unanue, E. R., Studies in listeriosis show the strongsymbiosis between the innate cellular system and the T-cell response. Immunol. Rev. 1997. 158: 11–25.

5 Mackaness, G., Cellular resistance to infection. J. Exp.Med. 1962. 116: 381–406.

6 Bancroft, G. J., Schreiber, R. D., Bosma, G. C.,Bosma, M. J. and Unanue, E. R., A T-cell-independentmechanism of macrophage activation by interferon- + . J.Immunol. 1987. 139: 1104–1107.

7 Bancroft, G. J., Sheehan, K. C., Schreiber, R. D. andUnanue, E. R., Tumor necrosis factor is involved in the Tcell-independent pathway of macrophage activation inscid mice. J. Immunol. 1989. 143: 127–130.

8 Kato, K., Nakane, A., Minagawa, T., Kasai, N., Yama-moto, K., Sato, N. and Tsuruoka, N., Human tumornecrosis factor increases the resistance against Listeriainfection in mice. Med. Microbiol. Immunol. (Berl.) 1989.178: 337–346.

9 Roll, J. T., Young, K. M., Kurtz, R. S. and Czuprynski,C. J., Human rTNF alpha augments anti-bacterial resis-tance in mice: potentiation of its effects by recombinanthuman rIL-1 alpha. Immunology 1990. 69: 316–322.

10 Havell, E. A., Evidence that tumor necrosis factor has animportant role in antibacterial resistance. J. Immunol.1989. 143: 2894–2899.

11 Nakane, A., Minagawa, T., Kohanawa, M., Chen, Y.,Sato, H., Moriyama, M. and Tsuruoka, N., Interactionsbetween endogenous gamma interferon and tumornecrosis factor in host resistance against primary andsecondary Listeria monocytogenes infections. Infect.Immun. 1989. 57: 3331–3337.


12 Liu, W. and Kurlander, R. J., Analysis of the interrela-tionship between IL-12, TNF-alpha, and IFN-gammaproduction during murine listeriosis. Cell. Immunol.1995. 163: 260–267.

13 Huang, S., Hendriks, W., Althage, A., Hemmi, S.,Bluethmann, H., Kamijo, R., Vilcek, J., Zinkernagel, R.M. and Aguet, M., Immune response in mice that lackthe interferon-gamma receptor. Science 1993. 259:1742–1745.

14 Rothe, J., Lesslauer, W., Lotscher, H., Lang, Y., Koe-bel, P., Kontgen, F., Althage, A., Zinkernagel, R.,Steinmetz, M. and Bluethmann, H., Mice lacking thetumour necrosis factor receptor 1 are resistant to TNF-mediated toxicity but highly susceptible to infection byListeria monocytogenes. Nature 1993. 364: 798–802.

15 Eugster, H., Müller, M., Karrer, U., Car, B. D., Schny-der, B., Eng, V. M., Woerly, G., Le Hir, M., di Padova, F.,Aguet, M., Zinkernagel, R., Bluethmann, H. and Ryf-fel, B., Multiple immune abnormalities in tumor necrosisfactor and lymphotoxin- § double-deficient mice. Int.Immunol. 1996. 8: 23–36.

16 Fehr, T., Schoedon, G., Odermatt, B., Holtschke, T.,Schneemann, M., Bachmann, M. F., Mak, T. W.,Horak, I. and Zinkernagel, R. M., Crucial role of inter-feron consensus sequence binding protein, but neitherof interferon regulatory factor 1 nor of nitric oxide syn-thesis for protection against murine listeriosis. J. Exp.Med. 1997. 185: 921–931.

17 Endres, R., Luz, A., Schulze, H., Neubauer, H., Futte-rer, A., Holland, S. M., Wagner, H. and Pfeffer, K., Lis-teriosis in p47(phox−/−) and TRp55−/− mice: protectiondespite absence of ROI and susceptibility despite pres-ence of RNI. Immunity 1997. 7: 419–432.

18 Bogdan, C., Vodovotz, Y., Paik, J., Xie, Q. W. andNathan, C., Traces of bacterial lipopolysaccharide sup-press IFN-gamma-induced nitric oxide synthase geneexpression in primary mouse macrophages. J. Immunol.1993. 151: 301–309.

19 Beckerman, K. P., Rogers, H. W., Corbett, J. A.,Schreiber, R., McDaniel, M. L. and Unanue, E.R.,Release of nitric oxide during the T cell-independentpathway of macrophage activation. Its role in resistancein Listeria monocytogenesis. J. Immunol. 1993. 150:888–895.

20 Bermudez, L. E., Differential mechanisms of intracellularkilling of Mycobacterium avium and Listeria monocyto-genes by activated human and murine macrophages.The role of nitric oxide. Clin. Exp. Immunol. 1993. 91:277–281.

21 Frei, K., Nadal, D., Pfister, H. W. and Fontana, A., Lis-teria meningitis: identification of a cerebrospinal fluidinhibitor of macrophage listericidal function as interleu-kin 10. J. Exp. Med. 1993. 178: 1255–1261.

22 Boockvar, K.S., Granger, D. L., Poston, R. M., May-bodi, M., Washington, M. K., Hibbs, J. J. and Kurlan-der, R. L., Nitric oxide produced during murine listeriosisis protective. Infect. Immun. 1994. 62: 1089–1100.

23 MacMicking, J. D., Nathan, C., Hom, G., Chartrain, N.,Fletcher, D. S., Trumbauer, M., Stevens, K., Xie, Q. W.,Sokol, K. and Hutchinson, N., Altered responses tobacterial infection and endotoxic shock in mice lackinginducible nitric oxide synthase [published erratumappears in: Cell 1995 81: following 1170]. Cell 1995. 81:641–650.

24 Kamijo, R., Harada, H., Matsuyama, T., Bosland, M.,Gerecitano, J., Shapiro, D., Le, J., Koh, S. I., Kimura,T. and Green, S. J., Requirement for transcription factorIRF-1 in NO synthase induction in macrophages. Sci-ence 1994. 263: 1612–1615.

25 Shiloh, M. U., MacMicking, J. D., Nicholson, S.,Brause, J. E., Potter, S., Marino, M., Fang, F., Dinauer,M. and Nathan, C., Phenotype of mice and macro-phages deficient in both phagocyte oxidase and induc-ible nitric oxide synthase. Immunity 1999. 10: 29–38.

26 Goossens, V., Grooten, J., De-Vos, K. and Fiers, W.,Direct evidence for tumor necrosis factor-induced mito-chondrial reactive oxygen intermediates and theirinvolvement in cytotoxicity. Proc. Natl. Acad. Sci. USA1995. 92: 8115–8119.

27 Ding, A. H., Nathan, C. F. and Stuehr, D. J., Release ofreactive nitrogen intermediates and reactive oxygenintermediates from mouse peritoneal macrophages.Comparison of activating cytokines and evidence forindependent production. J. Immunol. 1988. 141:2407–2412.

28 Phillips, W. A. and Hamilton, J. A., Phorbol ester-stimulated superoxide production by murine bonemarrow-derived macrophages requires preexposure tocytokines. J. Immunol. 1989. 142: 2445–2449.

29 Bermudez, L. E. and Young, L. S., Tumor necrosis fac-tor, alone or in combination with IL-2, but not IFN-gamma, is associated with macrophage killing of Myco-bacterium avium complex. J. Immunol. 1988. 140:3006–3013.

30 Brunelli, L., Crow, J. P. and Beckman, J. S., The com-perative toxicity of nitric oxide and peroxynitrite toEscherichia coli. Arch. Biochem. Biophys. 1995. 316:327–334.

31 Gardner, P. R., Nguyen, D. D. and White, C. W., Super-oxide scavenging by Mn(II/III) tetrakis (I-methyl-4-pyridyl) porphyrin in mammalian cells. Arch. Biochem.Biophys. 1996. 325: 20–28.

32 Hooper, D. C., Spitsin, S., Kean, R. B., Champion, J.M., Dickson, G. M., Chaudhry, I. and Koprowski, H.,Uric acid, a natural scavenger of peroxynitrite, in experi-


mental allergic encephalomyelitis and multiple sclerosis.Proc. Natl. Acad. Sci. USA 1998. 95: 675–680.

33 Pfeffer, K., Matsuyama, T., Kundig, T., Wakeham, M.A., Kishihara, K., Shahinian, A., Wiegmann, K., Oha-shi, P. S., Kronke, M. and Mak, T. W., Mice deficient forthe 55 kd tumor necrosis factor receptor are resistant toendotoxic shock, yet succumb to L. monocytogenesinfection. Cell 1993. 74: 457–467.

34 Kaufmann, S. H. and Flesch, I. E., Cytokines in anti-bacterial resistance: possible applications for immuno-modulation. Lung 1990. 168: 1025–1032.

35 Bortolussi, R., Vandenbroucke-Grauls, C. M., van-Asbeck, B. S. and Verhoef, J., Relationship of bacterialgrowth phase to killing of Listeria monocytogenes byoxidative agents generated by neutrophils and enzymesystems. Infect. Immun. 1987. 55: 3197–3203.

36 Leenen, P. J., Canono, B. P., Drevets, D. A., Voerman,J. S. and Campbell, P. A., TNF-alpha and IFN-gammastimulate a macrophage precursor cell line to kill Listeriamonocytogenes in a nitric oxide-independent manner. J.Immunol. 1994. 153: 5141–5147.

37 Krakauer, T. and Oppenheim, J. J., IL-1 and tumornecrosis factor-alpha each up-regulate both the expres-sion of IFN-gamma receptors and enhance IFN-gamma-induced HLA-DR expression on human monocytes anda human monocyte cell line (THP-1). J. Immunol. 1993.150: 1205–1211.

38 Chomarat, P., Rissoan, M. C., Banchereau, J. andMiossec, P., Interferon gamma inhibits interleukin 10production by monocytes. J. Exp. Med. 1993. 177:523–527.

39 Frankova, D. and Zidek, Z., IFN-gamma-induced TNF-alpha is a prerequisite for in vitro production of nitricoxide generated in murine peritoneal macrophages byIFN-gamma. Eur. J. Immunol. 1998. 28: 838–843.

40 Hauser, T., Frei, K., Zinkernagel, R. M. and Leist, T. P.,Role of tumor necrosis factor in Listeria resistance ofnude mice. Med. Microbiol. Immunol. 1990. 179:95–104.

41 Kelso, A., Glasebrook, O., Kanagawa, O. and Brun-ner, K. T., Production of macrophage-activating factorby T lymphocyte clones correlation with other lympho-kine activities. J. Immunol. 1982. 129: 550–556.

42 Hibbs, J. B., Taintor, R. R. and Vavrin, Z., Macrophagecytotoxicity: role for L-arginine deiminase and iminonitrogen oxidation to nitrite. Science 1987. 235:473–476.

43 Green, L. C., Wagner, D. D. A., Glogowski, J., Skep-per, P. L., Wishnok, J. S. and Tannenbaum, S. R., Anal-ysis of nitrate, nitrite and 15N-nitrate in biological fluids.Anal. Biochem. 1982. 126: 131–138.

Correspondence: Hans-Pietro Eugster, Section of ClinicalImmunology, Department of Internal Medicine, UniversityHospital Zürich, Häldeliweg 4, CH-8044 Zürich, SwitzerlandFax: +41-1-634 2902e-mail: hanspietro.eugster — dim.usz.ch


reduced antilisterial activity of tnf-deficient bone marrow-derived macrophages is due to impaired...

Documents