1354

Recombinant Murine Interferon-v Reversibly Activates Rat Alveolar Macrophages to Kill Legionella pneumophila Shawn J. Skerrett and Thomas R. Martin

Medical Research Service. VA Medical Center. and Division of Pulmonary and Critical Care Medicine. University of Washington School ofMedicine. Seattle

The interaction of interferon (IFN)-')', rat alveolar macrophages, and Legione/la pneumophila was studied in vitro to define the effector cell potential of alveolar macrophages against an intracellular pathogen in a model in which the efficacy ofIFN-')' could be tested in vivo.Alveolar macrophages preincubatedwith IFN-')' up-regulatedIa antigen and killed 0.5-4 logs of L. pneumophila over 4 days compared with 1-2 logs of bacterial growth in untreated cells. The bactericidal effect was dose dependent, evident over a wide range of bacterial inocula, and not suppressed by hydrocortisone. Preincubation with IFN-')' was unnecessary and insufficient, as intracellular replicationwasreversedbyexposureto IFN-')'up to 48 h after infection,and neutralization ofIFN-')' after infection permitted bacterialgrowth. IFN-')'thus convertsalveolar macrophages from target cells to effector cells in host defense against L. pneumophila and may be of therapeutic benefit in legionellosis.

Legionel/a pneumophila is the etiologic agent of legionnaires' disease and a facultative intracellular parasite of mononuclear phagocytes [1, 2]. The target cell of inhaled L. pneumophila is the alveolar macrophage, the resident phagocytic cell of the lung. L. pneumophila replicates within alveolar macrophages in vitro, and histologic studies have shown the pathogen to be located predominantly within alveolar macrophages in vivo [3-6]. The development of cell-mediated immunity leading to activation of macrophages to resist L. pneumophila is probably an essential host response to legionellosis. Legionnaires' disease is more common and more severe when cell-mediated immunity is impaired, particularly in the setting of corticosteroid therapy [7, 8]. Lymphocyte sensitization to L. pneumophila antigen has been demonstrated in humans and laboratory animals recovering from legionellosis [8-12], and lymphokine-activated mononuclear phagocytes inhibit the replication of L. pneumophila in vitro [10, 13-15]. Interferon (IFN)-'Y is the major macrophage-activating factor released by stimulated lymphocytes [16], and IFN-'Yactivated mononuclear phagocytes inhibit the growth of L. pneumophila in vitro [17-20]. The macrophage-activating

Received 3 February 1992; revised 28 July 1992. Presented in part: American Thoracic Society meeting. May 1991. Anaheim, CA (A399). Experiments were approved by the Animal Use Subcommittee of the Seattle VA Medical Center. Financial support: Merit Review and Career Development funds, Department of Veterans Affairs; grant AI-29IOJ, National Institutes of Health. Reprints or correspondence: Dr. Shawn J. Skerrett, General Medical Research (151). VA Medical Center. 1660 S. Columbian Way, Seattle. WA 98108. The Journal of Infectious Diseases 1992;166:1354-61 © 1992 by The University of Chicago. All rights reserved. 0022-1899/92/6606-0022$01.00

properties of IFN-'Y support a potential therapeutic role for this cytokine in the management ofintracellular infections in vivo, particularly in the compromised host, in whom IFN-'Y production is impaired. IFN-'Y has shown prophylactic. efficacy in animal models of a variety of intracellular infections and therapeutic potential in a limited spectrum of human disease [21]. The clinical use ofIFN-'Y is dependent on delivery of the cytokine to the target macrophage and requires a thorough understanding of the kinetics of the interaction of cytokine, target cell, and parasite. Although the feasibility of directly activating alveolar macrophages in vivo by systemic or airway administration ofIFN-'Y has recently been demonstrated [22-25], the kinetics of antimicrobial activation of alveolar macrophages by IFN-'Y are not fully understood, and the efficacy of IFN-'Y in a model of intracellular pulmonary infection has not been studied. We developed a rat model oflegionellosis to study pulmonary host defenses to intracellular pathogens and demonstrated that corticosteroid administration inhibits the clearance of inhaled L. pneumophila from the lung [8]. We recently reported that the resolution of legionellosis in normal rats is associated with the activation of alveolar exudate macrophages to express Ia antigen and inhibit the replication of L. pneumophila, supporting an effector cell function for immunologically activated alveolar macrophages in recovery from legionellosis [26]. We undertook the present studies to characterize the kinetics of the activation of alveolar macrophages by IFN-'Y to resist L. pneumophila in vitro, in a system in which the efficacy of IFN-'Y ultimately could be tested in vivo. Specifically, we sought to determine whether murine recombinant IFN-'Y activates rat alveolar macrophages to inhibit or kill L. pneumophila, whether pretreatment is required to observe this effect or whether established infection can be reversed by IFN-'Y, whether continued exposure to

JID 1992; 166 (December)

Alveolar Macrophages Kill L. pneumophila

IFN-'Y is required to sustain antimicrobial activity, and whether corticosteroids affect the responsiveness of alveolar macrophages to IFN-'Y.

Methods Bacteria. L. pneumophila Philadelphia I was originally obtained from American Type Culture Collection (ATCC 33152; Rockville, MD), passaged annually through rats, and stored as a lung homogenate [8, 26]. For each experiment, bacteria were thawed, grown in buffered charcoal yeast extract broth for 48 h, then incubated in buffered yeast extract broth for 1618 h as described [8, 26]. The bacteria were pelleted, washed twice in PBS, resuspended in PBS to an OD of 0.20-0.25 at 540 nm (-2 X 108 cfu/ml.), then diluted in RPMI 1640(Flow Laboratories, McLean, V A) as needed. Bacterial suspensions were quantitatively cultured on buffered charcoal yeast extract agar as described [8, 26] and the results expressed as colony forming units (cfu) per milliliter. Animals. Male Sprague-Dawley-derived rats weighing 180200 g and certified free of common pathogens were obtained from Simonsen Laboratories (Gilroy, CA). The rats were housed 2 or 3 per cage, allowed free access to food and water, and acclimated to the animal research facility for at least 4 days before study. Alveolar macrophages. Normal resident alveolar macrophages were obtained by bronchoalveolar lavage [26, 27]. Briefly, rats were anesthetized with pentobarbital and exsanguinated by cardiac puncture. The trachea was cannulated, the chest was opened, and the lungs were lavaged with four 10-mL aliquots of 0.85% sodium chloride containing 0.6 mM EDT A. The cells were washed twice in cold Hanks' balanced salt solution (HBSS) without calcium or magnesium (GIBCO Laboratories, Grand Island, NY) and resuspended in RPMI containing 5% heat-inactivated fetal calf serum (FCS) (Hyclone Laboratories, Logan, UT), penicillin, 100 units/ml., and streptomycin, 100 ~g/mL (GIBCO). The cells were counted in a hemocytometer, their viability was measured by the exclusion of trypan blue, and the differential was determined from cytocentrifuge specimens prepared with a modified Wright-Giemsa stain (DiffQuik; American Scientific Products, McGaw Park, IL). Alveolar cells prepared in this manner were >95% viable and >95% macrophages. IFN-'Y and anti-Il-Nrv. Recombinant murine IFN-'Y and rabbit polyclonal anti-murine IFN-'Y were gifts of Genentech (South San Francisco, CA). The IFN-'Y was supplied with a stated activity of 5 X 106 antiviral units/mg, I mg/ml., and an endotoxin level of < I0 pg/mL by the limulus amoebocyte lysate assay. The IFN-'Y was diluted in RPMI as needed and added to individual wells at the designated time in a volume of 10.0-12.5 ~L/well, yielding final concentrations of 10-1000 units/rnl., Anti-IFN-'Y was supplied as serum with neutralizing activity of 1.3 X 105 units of IFN-'Y/mL. The antiserum was diluted in RPMI to 4 X 104 neutralizing units/ml. and added to individual wells at the designated time in a volume of 12.5 ~L (500 units)/ well, for a final concentration of -1000 neutralizing units/rnl., Expression ofsurface Ia antigen by alveolar macrophages. To measure the effect of IFN-'Y on the expression of la antigen by

1355

rat alveolar macrophages, the cells were suspended in polypropylene tubes at I X 106 viable alveolar macrophages/ml, in RPMI, 5% heat inactivated FCS, penicillin, 100 units/ml., and streptomycin, 100 ~g/mL, with or without IFN-'Y, 250 units/ mL. After 20 h ofincubation at 37°C in humid air with 5%C02 , the cells were washed in HBSS and resuspended at - 3 X 106 alveolar macrophages/ml, in RPMI with 0.0 I% sodium azide. Surface la antigen was labeled by indirect fluorescence as described [26]. Briefly, the cells were incubated with the OX-6 murine monoclonal anti-rat la (diluted I: 100; SeraLab MAS 043, Westbury, NY) for 30 min at 4°C, washed twice in cold PBS, then incubated with phycoerythrin-conjugated goat antimouse IgG (diluted 1:40; Biomeda, Foster City, CA) for 30 min at 4°C. The cells were washed again, fixed in 2% paraformaldehyde, and studied by flow cytometry, using an Orthocytofluorograph model 50H with a model 2150 computer and model 164oI krypton laser (Ortho Diagnostic Systems, Westwood, MA) as described [26]. la-positive cells were defined as those exhibiting more fluorescence than 99%of the cells incubated with the second antibody alone. Injection ofalveolar macrophages. For each experiment, the alveolar macrophages from 6-8 rats were combined. The cells were diluted to I X 106 viable macrophages/ml, in RPM I with 5% heat-inactivated FCS, penicillin, and streptomycin, and 0.5mL aliquots of this suspension were added to 16-mm wells of 24-well clusters (Costar, Cambridge, MA). After 2 h of incubation at 37°C, the wells were washed six times with prewarmed HBSS to remove nonadherent cells and antibiotics. The monolayers and cell-free control wells then were covered with 0.5 mL of RPMI containing 10% fresh pooled rat serum and, in some experiments, hydrocortisone, 0.5-5.0 ~g/mL (Sigma, St. Louis, MO). IFN-'Y, anti-If-Nvv, or both were added to designated wells in 10.0-12.5 ~L of RPMI. After 20 h of incubation, each well was inoculated with L. pneumophila (5 X 104 to 5 X 107 cfu, estimated by optical density) in a volume of 20-25 ~L of RPM I or PBS, and the plates were mixed by gentle swirling. Immediately and at subsequent 24-h intervals, three or four wells representing each condition were subjected to ultrasonic oscillation to disrupt monolayers and release intracellular bacteria as described [26]. A 3.175-mm microtip attached to an ultrasonic cell disruptor was dipped in each well for 10 s. Initially, a Sonicator (Heat Systems Ultrasonics, Plainview, NY) with the output level set at 1.5 was used. A Branson Sonifier 250 (Branson Ultrasonics, Danbury, CT) adjusted to an output level of I and duty cycle of 60% was used in later studies. Exposure to ultrasonic oscillation under these conditions completely disrupted monolayers, as judged by inverted microscopy, but did not reduce the number of bacterial cfu. The sonicated suspensions were serially diluted in Mueller-Hinton broth and quantitatively cultured by spreading O.I-mL aliquots on buffered charcoal yeast extract agar with a bent glass rod. The number ofcolonies was counted after 3-4 days ofincubation at 37°C in 5%CO 2 , Data analysis. For studies ofla expression, each n represents alveolar macrophages from an individual animal. All other experiments were done with alveolar macrophages pooled from 6-8 animals, and each n represents a monolayer. Data are given as mean ± SE. Statistical comparisons were made using two-

Skerrett and Martin

1356

Media

...CD

..... Control -Ia

.c

E

::J

Z CD

o

...CD

.c

:

....

IFN-Y

·····Control -Ia

E

::J

Z Qj

o

10 100 Fluorescence Intensity

1000

Figure 1. Interferon-v(IFN-)') inducesla expression byalveolar macrophages. Top, resident alveolar macrophages incubated in culture media; bottom, cells incubated in media supplemented with IFN-)', 250 units/ml., la, cells incubated with mouse anti-rat la (OX-6), then with phycoerythrin-conjugated anti-mouse IgG; control, cellsincubatedwithsecondantibodyalone. Data are representative histograms from 3 individual animals.

tailed unpaired t tests. P < .05 was considered significant. The results of individual experiments are shown in each figure. Each experiment was done at least twice with similar results. All log scales are 10glO'

Results Recombinant murine fFN-)' induces fa expression by rat alveolar macrophages. To determine the species cross-reactivity of murine IFN-'Y with rat macrophages, we measured its effect on expression of the class II marker la. As shown in figure I, very little la was detected on normal resident alveolar macrophages incubated overnight in media. In contrast, cells incubated with IFN-'Y distributed into discrete Ia-positive and Ia-negative populations. Quantitatively, la was detected on 52.3% ± 2.9% of the alveolar macrophages incubated with IFN-'Y vs. 2.7% ± 0.6% of the control cells (mean ± SE, n = 3 rats, P < .000 I). Thus, exposure to IFN-'Y induces la expression by a subset of alveolar macrophages. fFN-'Y activates alveolar macrophages to kill L. pneumophila in a dose-dependent manner. Figure 2 illustrates that normal resident alveolar macrophages permitted L. pneumophila to replicate by - 2 logs over 3 days in culture. No replication occurred in wells containing media alone, suggesting that bacterial growth was intracellular. Preincubation of alveolar macrophages with 10 units/rnl, IFN-'Y partially inhibited replication of L. pneumophila. However, alveolar macrophages preincubated with 100 or 1000 units/ml, IFNI' not only inhibited bacterial growth but established net

lID 1992; 166 (December)

elimination of L. pneumophila over 3 days in culture, reducing the number ofcfu by > I log from the inoculum on day O. Similar experiments extended to 4 days in culture showed further net reduction in cfu. Bacterial killing was a very consistent finding: a net reduction in cfu of 0.5-4.0 logs over 3-4 days was observed in all 14 experiments in which alveolar macrophages were preincubated with ;;;.100 units/ml. IFN-'Y before challenge with L. pneumophila. IFN-'Y had no direct effect on bacterial viability in cell-free media (figure 2) and did not affect the viability of alveolar macrophages, determined by the exclusion of trypan blue after 4 days in culture (data not shown). Daily examination of culture plates by inverted microscopy revealed that IFN'Y-treated alveolar macrophages exhibited dose-related aggregation, and monolayers remained intact throughout the course of the experiment. Untreated infected monolayers showed no aggregation and were progressively destroyed over 4 days in culture. Thus, pretreatment ofalveolar macrophages with IFN-'Y caused a dose-related increase in resistance to L. pneumophila, and concentrations ofIFN-'Y ;;;.100 units/rnl. induced net bacterial killing. Therefore, we chose 250 units/rnl, for subsequent experiments to ensure maximum effect. fFN-'Y-activated alveolar macrophages kill L. pneumophila over a wide range ofbacterial inocula. As shown in figure 3, alveolar macrophages pretreated with IFN-'Y (250 units/ml.) resisted L. pneumophila at four different inocula spanning a 3-log range. Transient bacterial replication was seen under some conditions in the experiment shown and was observed at some inocula in one of two experiments identical to the one illustrated in figure 3. However, net elimination of L. pneumophila was established by 4 days in culture at each inoculum. Thus, the antibacterial resistance of IFN-'Y-

7-r-------------'-.--------, .-.~

.--.AM . - -.AM

6

5

.-.~

+ IFN 10 U / + IFN 100 U + IFN 1000 U

* * --

~

•

--

0··· OMedio -Media

o

..

*----_.:

»> ..

.i-> • •-

. .......~

4

* ~

~

~

~*

+ IFN 1000 U

-~*

*

2

3

Doys in culture

Figure 2. Interferon-v (IFN) activates alveolar macrophages (AM) to kill L. pneumophila in dose-dependentmanner. Data are mean ± SE of quadruplicate monolayersfor each conditionat each time point; vertical axis is 10glO scale. *, P < .01 compared with media on same day. Representative of three experiments.

Alveolar Macrophages Kill L. pneumophila

JID 1992; 166 (December)

.-.AM

7 ........ Ol

6

0

.;::, ...J

E

u,

o

5 4

~.

~~~; !~-------.~ • • . . . . .(> . • . . . • . •(> ., ..•.•

........ Ol

0 .;::,

IFN T-l IFN TO ~-~AM IFN Tt IFN 1'2 0··· OMedla

'0

u,

*

~~~t

4 ".~~:"'\-

*

E

t

,'3

5

.-.AM .--.AM .--.AM +++ .-.AM +

...J

~.

.

6

+ IFN

O"'OMedia

1357

~-....

,'3

......

(,)

..."~~~ .....

,

~

-o ..... Q

...... * '

*

't---~~

::-,:-....

2

.

'i ~

o

2

,'3

4

Days in culture

treated alveolar macrophages was effective over a broad range of bacteria-to-cell ratios. Antibacterial activity of IFN-"(-treated alveolar macrophages is not augmented by daily supplementation of IFN"(. To determine if further antibacterial activity could be induced in alveolar macrophages that had been treated with IFN-"( for 20 h before challenge with L. pneumophila, we tested the effect of adding fresh IFN-"( to the media each day. As shown in figure 4. alveolar macro phages pulsed with daily IFN-"( did not exhibit greater resistance to L. pneumophila than the cells treated with a single dose ofIFN-"(. IFN-"( reverses established intracellular infection with L. pneumophila. To determine whether pretreatment with

........

5

.-.AM .--.AM •-.AM

0·· . o MedIa

.

Ol

0

.;::,

,

+ IFN + IFN QD

.

·················0

4

...J

E