Immunology Letters, 27 ~1991) 237- 242 Else*ier

IMLET 01547

Identification of cytokines which enhance (CSF-1, IL-3) or restrict (IFN-3,) growth of intramacrophage Listeria monocytogenes

Michel Denis I and Evan O. Gregg ~ I(-"ttitO de Recherche, Centre de Pneumologie, Hdpital Laval, Samte-Foy, Canada and "Bioscience I, ICI Pharmaceuticals,

Mereside, A lderlev Park, Macclesfield, Cheshire, t..~K.

(Received 8 October 1990; re,,ision received 29 No~ember 1990; accepted 2 December 19901

1. Summary

Murine peritoneal macrophages were isolated by adherence and their listericidal activity assessed in the presence or absence of selected cytokines. Un- treated macrophages were not highly listericidal, showing moderate killing in the first 2 h after infec- tion, and allowed progressive microbial growth thereafter (up to 9 h). Pre-treatment of cells with 10 to 100 U/ml of IFN-3, allowed macrophages to de- velop sustained listericidal activity for the 9-h obser- vation period, with a 2-log reduction of Listeria CFU per monolayer. Pulsing of cells with TNF-a alone did not result in enhanced microbicidal activi- ty but TNF-o~ potentiated IFN-',/-induced listericidal activity, resulting in high levels of killing when both cytokines were present. Conversely, macrophages pre-treated with interleukin-3 (IL-3) or colony- stimulating factor-I (CSF-I) were found to be much more permissive for Listeria growth. Neither IL-3 nor CSF-I abrogated IFN--~-induced listericidal ac- tivity. Moreover, neither IL-3 nor CSF-I had any ef- fect on the ability of macrophages to develop a respiratory burst following Listeria infection, as judged by H202 release following in vitro infection. Overall, these results suggest that different cytokines may have opposing effects on intracellular microbial growth, and that the balance of c.vtokine production

Key words.. Intracellular pathogen; Macrophage; Cytokine

Correspondence to: Michel Denis, Unite de Recherche, Centre de Pneumologie, H6pital Laval, 2725 Chemin Sainte-Foy, Sainte- Foy. Canada, GIV 4G5.

in vivo may determine the resistance or susceptibility of the infected host.

2. Introduction

Modulation of macrophage bactericidal activity during infection with intracellular pathogens is an area of intense study. Murine listeriosis has been studied extensively to develop our understanding of infections with virulent microbes [1]. Investigations in this model have allowed descriptions of the impor- tance of elements such as the genetic makeup [21, T- cell response [31, phagocytic cell activation [41 and cytokine secretion [5] in determining resistance to Listeria monoo'togenes.

During listeriosis, it is believed that macrophages are important effector cells in killing microbes in the infected organs. This is supported by the fact that "activated" macrophages in vivo [61 or in vitro [71 are highly listericidal, and that elimination of mac- rophages in vivo by treatment with a monoclonal an- tibody against a complement receptor greatly ex- acerbates Listeria infection by allowing parenchymal cells to become infected, with dramatic consequent bacterial multiplication [8]. However, it is still not entirely clear what characterises a "listeri- eidal" compared to a "resting" macrophage, and how various soluble factors mediate phenotypic changes in the macrophages, including bactericidal activity.

Recent evidence has accumulated to show that interferon-gamma (IFN-~), a T-cell product, is a po- tent modulator of macrophage effector function [91. Although it is apparent that depletion of IFN-'~ in

0165-2478 91 $ 3.50 '~' 1991 Elsevier Science Publishers B3,'. 237

vivo leads to greater susceptibility to listeriosis [10], it is still a matter of debate if IFN-7 induces macro- phage listericidal activity. Some reports suggest a beneficial effect [11] whereas other workers find IFN--~-pulsed cells no more bactericidal than control phagocytes [12]. Also not clear is the effect of cytokines, other than IFN-3,, on macrophage hand- ling of Listeria. This is an important issue as the cytokine profile elicited during Listeria infection in- cludes IFN-% interleukin-2 (IL-2), TNF-~ [13] as well as colony stimulating factors [14]. This study ex- amines how these cytokines modulate listericidal ac- tivity of mouse macrophages.

3. Materials and Methods

3.1. Bacterial strain

Listeria monocytogenes strain EGD was main- tained by frequent inoculation in BALB/c mice, to retain its virulence, and plating of spleen homogenate on tryptic soy agar (Difco, Detroit, MI). Before use in the killing assays, L. monocytogenes was grown for 10 h to log phase in trypticase soy broth (Difco). Bacteria were then cen- trifuged at 4000xg for 10 min, washed, suspended in saline and frozen in aliquots at - 7 0 ° C . Before use in killing assays, bacteria were thawed and brief- ly sonicated (5 s).

3.2. Cytokines

Murine IFN-7 was obtained from Genentech (South San Francisco, CA) at a specific activity of 1.9x 10 7 U/mg. Mouse TNF-ot (Genentech) had a specific activity of 2.9xl07 U/mg. Mouse CSF-I and mouse IL-3 were both obtained from Amgen (Boston, MA), the specific activity was 2.1 × 106 U/rag and 3.1 x 106 U/mg, respectively. All cytokines were diluted in pathogen-free saline.

3.3. Peritoneal macrophages

Specific pathogen-free BALB/c mice bred in our facility were used throughout. Peritoneal exudates from untreated mice were obtained by washing the peritoneum with l0 ml of ice-cold RPMI-1640 (Gib- co, Grand Island, NY). Washes from groups of 5 - l0 mice were pooled. Peritoneal cells were washed once

238

in RPMi, and adjusted to 10 6 ceils/ml and plated at 5 ml on 12-mm cover slips (Nunc, Roskilde, Den- mark) in petri dishes (Nunc), in RPMI-1640 fortified with 2 mM glutamine (Gibco) and 10°7,0 foetal calf serum (Gibco). Dishes were placed in a 5°7o CO2/95% humidified air atmosphere for 2 hours to allow adherence. After this step, non-adherent cells were removed by vigorous washings with warm medium. Cells were then re-incubated for 18 hours in the presence or absence of indicated amounts of cytokines before infection with Listeria. Before in- fection, cells were more than 9507o macrophages, as judged by Giemsa staining.

3.4. Infection in vitro

Monolayers of murine peritoneal macrophages were infected with 2× 10 6 bacteria/ml for 30 min. After this phagocytosis period, cover slips were washed vigorously with warm medium and were reincubated in a 507o CO2/95% humid air at- mosphere. Immediately (to), and at pre-determined intervals thereafter, bacteria associated with monolayers were assayed microbiologically as described below. Sample coverslips, in triplicate, were taken and deposited into plastic tubes contain- ing 5 ml distilled water. Cells were lysed by vigorous mixing and a brief sonication for 10 s, which, in pi- lot experiments, was found to be innocuous for Listeria. Serial dilutions of these lysates were manu- ally plated on BHI agar (Difco), three plates were used per coverslip. The number of bacteria in the su- pernatants was also determined by plating on agar and this never exceeded 5% of the cell-associated bacteria at all time points shown in the Results. Data for all time points in the Figures represent an average of these triplicate determinations, for one represen- tative experiment. All experiments were repeated on at least 3 occasions with similar results.

3.5. Release of hydrogen peroxide

H202 release by BALB/c murine peritoneal mac- rophages was determined by the horseradish peroxidase-mediated, H202-dependent oxidation of homovanillic acid as described [15]. Briefly, 3-ml ali- quots of a suspension containing l06 peritoneal macrophages were incubated at 37 °C in 6-ml plastic tubes (Nunc) containing l unit of horseradish perox-

i[lase/ml (Sigma, St. Louis, MO) and 100 #M homovanillic acid (Fluka, Switzerland). Cells were stimulated with either 100 ng/ml of phorbol- myristic acetate (PMA) (Sigma) or 107 L. monocytogenes. After 1 h, the reaction was stopped by adding 0.25 ml of 0.1 M glycine/NaOH buffer (pH 12) with 25 mM EDTA. After centrifugation, the homovanillic reaction product in the superna- tant was determined fluorometrically with a Perkin- Elmer model 3000 spectrofluorimeter (Verberlin- gen, F.R.G.) as a measure of the amount of H202 released. Excitation was at 312 nm [15]. The release of H202 was expressed as nmol. H202/106 ceils/60 min.

3.6. Statistical analysis

Differences between experimental means were analysed by Student's t-test.

4. Results

L. monocytogenes strain EGD successfully infect- ed resident peritoneal macrophages from BALB/c mice as seen from bacterial uptake (Fig. 1: bacterial numbers at 30 min). Both untreated and IFN-3,- pulsed macrophages were listericidai for the first

1.5 h and a 10-fold decrease in bacterial CFU was observed at 2 h post-phagocytosis (Fig. IA). In un- treated macrophages, surviving Listeria started dividing slowly and by 9 h after phagocytosis bac- terial numbers in untreated monolayers were similar to those seen following phagocytosis. It is important to note that this increase was not due to lysis of mac- rophages and release of bacteria in the permissive ex- tracellular environment as: (i) bacterial numbers in the extracellular medium were always low (5°'/0 of that cell-associated); and (ii) cell number did not diminish significantly during in vitro infection (data not shown).

IFN-y-treated macrophages were listericidal, and did not allow bacterial multiplication after the initial killing burst such that a 2-log CFU reduction was seen at the end of the observation period (Fig. IA). Also shown in Fig. IA is that 100 U/ml or 10 U/ml of IFN-7-induced potent killing activity. TNF-a alone induced no killing activity but it potentiated IFN-~-induced listericidal activity, with a 0.8-log CFU increase in killing at 2 h and sustained lower CFU counts in TNF-ot/IFN-~-pulsed monolayers vs. IFN-y alone treated (Fig. IB).

Colony stimulating factors such as CSF-I, GM- CSF and/or I L-3 have been shown to increase macro- phage effector functions against turnouts and mi-

A

6 6

s' s

~ 4 o 4

i i I I I I I I I I

0 1 2 3 4 5 6 7 8 9

Hours

B

• ' W A ' J

'A'

I I 2 3 4 5 6 7 8 9

Hours

Fig. I. Growth o f Listeria in BALB/c m o u s e peri toneal macrophages . (A) Macrophages were left unt rea ted ( • ), pulsed overnight with 10 U / m l ( o ) or 100 U /ml ( • ) o f [FN-7 and their listericidal activity assessed as descr ibed in Methods . (B) Macrophages were left unt rea ted ( • ), pulsed with I000 U / m l T N F - a ( o ), pulsed with 1000 U /m [ IFN-7 ( • ) or IO00 U / m l TNF-~ and IFN-~ ( • ) and their killing acti~it~ assessed. Each point represents the m ean of three cover slips. SEM did not exceed 15% of the mean values and are omi t ted for clarity.

* Denotes P < O . 0 5 vs. all o ther exper imenta l groups.

239

crobes [16, 17]. Accordingly, we pulsed our monolay- ers with CSF-! and IL-3 to examine the effect o f these cytokines on microbic ida l activity. The results in Fig. 2 indicate that CSF-I or IL-3 t rea tment rendered cells more permissive o f Listeria growth a l though this t rea tment did not affect the initial kil l ing burst. Presence o f CSF-1 or IL-3 led to a one- log increase in C F U counts per monolayer at 9 h af ter infect ion ( P < 0,0001). Cell enumera t ion showed that there was no s ignif icant decrease or increase in cell uptake dur ing infect ion (da ta not shown).

In repor ted studies on mac rophage kill ing o f the in t racel lu lar paras i te Leishmania major, it has be- come apparen t that M A F activity, ascr ibed to IFN- % can be abroga ted by inclusion o f o ther cytokines, no tab ly IL-3 and IL-4 [18]. We therefore sought to investigate whether this impa i rmen t o f microbic ida l act ivi ty also app l ied to our Listeria system. The da ta ob ta ined , summar i sed in Fig. 3, suggest that CSF-I a lone d rama t i ca l ly increased mac rophage permis- skeness to Listeria with an exponent ia l growth o f the microbe between 3 and 9 h pos t -phagocytos is . S imi lar results were ob ta ined for I1.-3 (da ta not shown). IFN-~-pu l sed mac rophages were l istericidal

9 -

8 -

i 7 - 6

3 I I I I I I I I I

1 2 3 4 5 6 7 8 9

Hours

Fig. 3. Gro,~th of Ltsterta in BALB c peritoneal macrophage~ x~hich ~ere left untreated I, • ), pulsed overnight ~ith 1000 U ml CSF-I I • k 1000 U ml of IFN-) ( .a ) or 1000 U ml CSF-I and IFN-5 ( o ). Results are expressed as m Fig. I. SEM did not exceed 10% of the mean ~alues and are omitted for clarity. * Denote~

P<0.01 ~s. all other experimental groups.

._o.

6 >

8 . m

o

" 5 O

g~ . .A

7

4 I I I I I I I I I

0 1 2 3 4 5 6 7 8 9

Hours

Fig. 2. Growth of Lisleria in BALB,.c peritoneal macrophages which ~ere left untreated ( • ), pulsed with 1000 U.,ml IL-3 ( t~ ) or 1000 U/ml CSF-I (o). Results are expressed as in Fig. I. SEM did not exceed 15% of the mean values and are omitted Ior clarity.

* Denotes P<0.OI ~s. all other experimental groups.

and did not al low microbia l growth dur ing the course o f the exper iment , and inclusion o f CSF-I (Fig. 3) or II.-3 (da ta not shown) did not mod i fy this growth pat tern. Thus, the beneficial effect o f IFN-~ is not abroga ted by t rea tment with ei ther !I_-3 or CSF-I .

We sought to de te rmine the effect o f these cytokines on an impor tan t effector funct ion o f mac- rophages, the respi ra tory burst , which is involved in the kill ing o f a variety o f microbes /paras i tes . At present, the exact role o f reactive oxygen in termedi- ates (ROI) in the kil l ing o f Listeria by act ivated mac- rophages is unclear [19]. The H20_, release o f resi- dent per i toneal macrophages upon tr iggering with P M A or Listeria was therefore examined when cells were pulsed with various cytokines (Table 1). Our da ta show that both P M A and Listeria t r iggered a subs tant ia l release o f H20_, from macrophages . Pulsing with i000 U/ml o f IFN-"t. for 18 h led to mac rophages which developed a s igni f icant ly higher respi ra tory burst (approx. 2-fold increase) af ter ei ther s t imuli ( P < 0.01). App l i ca t ion o f IL-3 or CSF- 1 with IFN- 3, d id not mod i fy this release potent ia l .

240

TABLE 1

H_,O, release b.,, resting and peritoneal macrophages stimulated ~lth PMA or Listerta x~as measured as the horseradish peroxidase-mediated H20_,-dependent oxidation of homo~anil-

lic acid. Values are the means of 5 experiments -+- standard devia- tions.

Treatment a H~O, release (nmol 10 (' cells h)

Resting PMA h Lm c

A None <0.3 2.8+0.7 2.0--0.4 B IFN-') (1000 U ml) <0.3 6.3+ 1.2 d 4.6_+ 1.0 d C IL-3 (1000 U ml) <0.3 2.7_+0.5 1.9+_0.5 D CSF-I (1000 U.'ml) <0.3 2.6+0.8 2.4_+0.7 E IFN--~ _+ IL-3

(both at I000 U ml) <0.3 6.6-+ 1.3 d 5.0± 1.3 d F IFN--) _+ CSF-I

(both at I000 U..ml) <0.3 6.0+_0.90 4.5___0.9 d

aBALB, c peritoneal macrophages were pulsed for 18 h before measurement of H,O, release under xarious conditions, bFinal concentration of PMA was IOOng,.ml. ~10 ~ cfu of L. monoc.vtogenes were added to 10 ° macrophages, dp< 0.01 ~er-

sus untreated cells.

Moreover, cells pulsed with IL-3 or CSF-I alone secreted the same amoun t o f H20 2 as untreated ceils when triggered with P M A or Listeria. These results show that CSF-1 or IL-3 treatment did not render macrophages more permissive for microbial growth by impairing the activity o f the respiratory burst.

5. Discussion

Macrophage activity and its modula t ion by cytokines is a central issue in our understanding of host resistance to intracellular pathogens. Murine listeriosis induces activated macrophages in the or- gans o f infected mice [20] and it is believed that these cells are the final effectors against invading Listeria. Products o f sensitised T lymphocytes are also im- por tant for an efficient macrophage listericidal re- sponse [21]. Indeed, administrat ion o f antibodies directed against IFN- ) greatly enhanced mouse sus- ceptibility to Listeria [10] and direct administrat ion o f IFN-2; increased resistance [22].

Our studies have focused on the direct interaction between macrophages and L. monocytogenes and its modula t ion by cytokines. From our results (ref. 23 and this study) it is not surprising that earlier studies

showed a lack o f effect o f IFN-~, on listericidal ac- tivity at an early time period ( 1 - 2 h) [12, 24]. In our system, IFN--),-induced listericidal activity takes longer to develop and occurred after an initial cytokine-independent killing burst, in all macro- phages (Figs. 1 -3) . Thus, it appears that IFN--) acts by inducing a sustained programme of bactericidal activity for at least a 9-h period. Moreover, IFN-') induced a 2- fold increase in respiratory burst activity in macrophages, triggered with P M A or Listeria (Ta- ble 1) a l though we do not know if the increased bac- terial killing was linked to this superior burst.

Recent results with human monocytes show that IFN-3, increased killing o f Listeria and that this kill- ing was abrogated by adding superoxide dismutase [25]. Other workers have suggested the involvement o f reactive nitrogen intermediates, namely nitric ox- ide, in IFN-) - induced killing o f other microbial spe- cies [26] and we are pursuing these possibilities in the Lister& system. Nevertheless, SOD seems to be a virulence factor for Listeria and some other microorganisms [27], whereas other ROI scavengers like catalase are not [28].

It is now quite clear that the enhanced product ion o f a variety o f c.vtokines, notably TNF-a, IL-2 [13] and CSFs [141 is seen in the course o f this disease. It is therefore essential to understand the activity o f these factors on macrophage listericidal activity to dissect their role in host resistance. In our hands, TNF-o~ itself did not increase the listericidal activity o f macrophages but it did, however, increase IFN--y- induced killing activity. These findings strongly sug- gest that the product ion o f TNF-o~ and IFN-3, may have a profound beneficial effect in host resistance. Indeed, depletion o f IFN--) or TNF-c¢ leads to great- ly increased susceptibility o f the infected host [10, 291.

Many CSFs are potentially present in the serum and the organs o f infected mice but CSF-I seems to predominate in listeriosis [14]. In our system neither CSF-I nor Ik-3 influenced the early phase o f killing (first 2 h), whereas they did increase subsequent growth of the surviving bacteria. In this phase, growth of Listeria is exponential and the doubling time o f the microbe is approximately 40 min (see Fig. 3). These results seem at odds with those o f Cheers et al. [30] who found that CSF-l-treated mac- rophages phagocytosed more bacteria whereas the bactericidal activity was unaffected. However, these

241

authors examined bacteriolytic activity at 2.5 h after uptake, a time at which killing by macrophages is un- affected by all soluble factors which we have tested. We have not verified the reported CSF-l-stimulated increase in Listeria uptake in our study. However, we use plating as a measure of the infection, a method which is relatively inaccurate when compared to the isotope release system used by Cheers et al. [30].

In our study CSF-I and IL-3 enhanced macro- phage permissiveness of Listeria growth whereas they did not modify IFN-7-induced listericidal ac- tivity. This is in contrast to leishmaniacidal activity of mouse macrophages where IFN-7-activated kill- ing was abrogated by IL-3 or GM-CSF in L. major- infected mice [31]. In all of these cases the mecha- nism involved in cytokine-induced enhanced microbial growth is not known.

From our results it is tempting to suggest that secretion of IFN-7/TNF-ot is associated with resist- ance to Listeria, whereas CSFs may be associated with susceptibility. The situation is likely to be more complex in an infectious process where there are many interactions between different factors. Howev- er, recent data clearly show that exogenous adminis- tration of IL-2 or IFN-7 enhanced host resistance to Listeria and also diminished CSFs released in the se- rum of protected animals [32, 33]. In other studies GM-CSF administration has been shown to increase host resistance against Listeria when infused directly into infected mice [34]. The quantity injected, how- ever, was quite high (4 g/mouse/day) and may imply that GM-CSF was acting by inducing another factor rather than by a direct action on macrophages. Col- lectively, these recent studies suggest that exogenous administration of cytokines may enhance host resist- ance against infectious agents. However, before this is extended to the clinical setting, a careful descrip- tion of the effect of cytokines on bacterial growth is warranted.

References

Il l North, R. J. 0983) J. Exp. Med. 138, 342. [2] Czuprynski, C. J. and Brown, J. E (1986) Immunology 58,

437. [3l Berch, P., Decreusefond, C., Theodorou, I. and Stiffel, C.

(1989) J. Immunol. 132, 932.

[4] Ohara, R.. Mitsuyama, M., Mi~ata, M. and Nomoto, K. (1985) Infect. lmmun. 48, 763.

[5] Bancroft, G. J., Sheehan, K. C. F., Schreiber, R. D. and Unanue, E. R. (1989) J. [mmunol. 143, 127.

[6] Roesler, J., Grottlub, E., Baccarini, M. and Lohmann- Mathes, L. (1989) J. lmmunol. 143, 1710.

[7] Portnoy, D. A., Schreiber, R. D., Connelly, P. and Tilne.~, L. G. (1989) J. Exp. Med. 170, 2141.

[8] Rosen, H., Gordon, S. and North, R. J. 11989) J. Exp. Med. 170, 27.

[9] Schultz, R. M. and Kleinschmidt. ~,V. J. (1983) Nature 305, 239.

[10] Buchmeier, N. A. and Schreiber, R. D. (1985) Proc. Natl. Acad. Sci. USA 82, 7404.

[l l] Peck, R. (1985) J. Immunol. Methods 82, 7404. [12] Van Dissel, J. T., Stikkelbroeck, J. J. M., Michel, B. C., Van

Den Barselaar, M.T., Leigh, P. C. J. and Van Furth. R. (1987) J. Immunol. 139, 1673.

[ 13] Kratz, S. S. and Karlander, R. J. (1988) J. Immunol. 141,598. [14] Cheers, C., Haigh, A. M., Kelso, A., Metcalf, D., Stanle,,;

E. R. and Young, A. M. 0988) Infect. Immun. 56, 247. [15] Ruch, W., Cooper, P. H. and Baggiolini, M. (1983) J. lm-

munol. Methods 63, 347. [16l Karbassi, A., Becket, J. M., Foster. J. S. and Moore, R. N.

(1987) J. Immunol. 139, 417. [17J Lee, M. 1-. and Warren, M. K. 0987) J. Immunol. 138, 3019. [18] Lieu, F. Y., Millott, S., Li, Y., Lelchuk, R., Chart, W. L. and

Ziltener, H. 0989) Eur. J. Immunol. 19, 1227. [19] Godfr~; R. W. and Wilder, W. S. (1984) J. Leukoc. Biol. 36.

533. [20] Mackaness, G. B. (1969) J. Exp. Med. 129, 973. [2 I I Lane, F. C. and Unanue, E. R. (1972) J. Exp. Med. 132, 1104. [22] Kiderlen, A. E S., Kaufmann, S. H. E. and Lohmann-

Matthes, M. L. (1984) Eur. J. [mmunol. 14, 964. [23] Denis, M. and Gregg, E. O. (1990) Can. J. Microbiol. (in

press). [24] Campbell, P. A.. Conono, B. P. and Cook, J. L. 11988) In-

fect. Immun. 56, 1371. [25] Peck, R. 11989) J. Leukoc. Biol. 46, 434. [26] Ding, A. H., Nathan, C. F. and Stueho, D. J. 11988) J. Im-

munol. 141, 2407. [271 Welch, D. E, Sword, C. P., Brehm, S. and Dusanic, D. (1979)

Infect. Immun. 23, 863. [28] Leblond-Francillard, M., Gaillard, J. L. and Berche, P.

(1989) Infect. Immun. 57, 2569. [29] Havell, E. A. (1989) J. Immunol. 143, 2894. [30] Cheers, C., Hill, M., Haigh, A. M. and Stanley, E. R. 11989)

Infect. Immun. 57, 1512. [31] Liew, F. Y. (1989) Immunol. Today I0, 40. [32] Haak-Frendscho, M., Young, K. M. and Czuprynski, C. J.

(1989) Infect. lmmun. 57, 3014. [331 Kurtz, R. S., Young, K. M. and Czuprynski, C. J. (1989) In-

fect. lmmun. 57. 553. [34] Magee, D. M. and Wing, E. J. (1989) J. Immunol. 143, 2336.

242