Neutrophils: Response to Escherichia coli Endotoxin - Infection and ...

2 downloads 0 Views 1MB Size Report
Shandon Southern Products, Ltd., Runcorn, Cheshire, En- gland) stained with a ...... Dis. 127:845-850. 9. Falk, W., R. H. Goodwin, and F. J. Leonard. 1980. A 48- ...
Vol. 57, No. 3

INFECTION AND IMMUNITY, Mar. 1989, p. 810-816 0019-9567/89/030810-07$02.00/0 Copyright © 1989, American Society for Microbiology

Rat Alveolar Macrophage Production of Chemoattractants for Neutrophils: Response to Escherichia coli Endotoxin JOHN W. CHRISTMAN,lt* STEPHEN F.

PETRAS,1 PAMELA M. VACEK,2

AND GERALD S.

DAVIS'

Departments of Internal Medicine' and Biostatistics,2 University of Vermont, Burlington, Vermont 05405 Received 11 August 1988/Accepted 15 November 1988

Endotoxemia in rats is associated with the accumulation of neutrophils (polymorphonuclear leukocytes) within the airspaces of the lung. Polymorphonuclear leukocyte influx appears to be regulated by the intrapulmonary accumulation of chemotactic activity. Since alveolar macrophages (AMS) are prevalent cells in the airspace and are known to release a variety of chemotactic factors, we investigated the effect of endotoxin exposure on AM production of chemotactic activity. We tested the hypothesis that endotoxin-exposed AMs have an augmented ability to produce chemoattractants. We recovered AMs by bronchoalveolar lavage from control rats and from rats treated in vivo with a "low dose" (2.5 mg/kg) or a "high dose" (5.0 mg/kg) of Escherichia coli endotoxin. These AMs were then cultured in vitro for 15 h in the absence or the presence of endotoxin (15 and 30 ,ug/ml) to stimulate the cells to produce chemoattractants. We found that in vitro endotoxin stimulated normal AMs to secrete chemoattractants in a dose-dependent fashion. AMs from rats treated with endotoxin in vivo spontaneously secreted more chemoattractants than AMs from control rats. Exposure to in vivo endotoxin followed by in vitro stimulation with endotoxin resulted in an even greater production of chemoattractants by AMs. We found a significant association between the percent polymorphonuclear leukocytes recovered by bronchoalveolar lavage from the airspaces and the production of chemoattractants by AMs from the same specimen. The level of chemotactic activity spontaneously produced by AMs predicted the degree of stimulated production of chemotactic activity. Partial purification indicated that this chemotactic activity has two molecular weight peaks, one near 1,000 and the other near 50,000. The activity was stable at 100°C for at least 30 min and was degradable by trypsinization. We conclude that endotoxin can induce AM production of chemoattractants and that prior exposure to endotoxin in vivo affects the response of AM to in vitro endotoxin exposure. By inference, it is possible that this endotoxin-macrophage interaction may serve as a biologic amplifier of the effects of endotoxin and may have a role in the pathogenesis of septic lung injury in humans.

Many forms of acute lung injury in animals and humans are associated with the accumulation of neutrophils in the airspaces of the lung (4, 8, 26). The factors which initiate this inflammatory cell recruitment have not been completely defined. Granulocyte movement into the lung, initiated from the alveolar side of the alveolar-capillary barrier, is termed directed migration or chemotaxis and occurs in response to chemotactic factors. The alveolar macrophage (AM) is the source of a variety of chemotactic factors for neutrophils (20). Macrophages can elaborate and activate complement locally and produce grandients of C5a, a substance with potent chemotactic properties (25; J. E. Pennington, W. J. Matthews, T. H. Rossing, D. J. Gash, F. S. Cole, and H. R. Colten, Clin. Res. 29:450A, 1981). Macrophages can generate interleukin-1 with many postulated effects, one of which is the mobilization of marginated pools of neutrophils (7). Finally, the macrophage can elaborate peptides and eicosinoids (leukotriene-B4) which may serve as chemoattractants (10, 16, 17). Endotoxins are lipopolysaccharide-protein complexes contained in the cell walls of gram-negative bacteria and other microorganisms. Endotoxin is capable of triggering most aspects of the acute inflammatory response, including activation of macrophages (19). When injected into the bloodstream or peritoneum of a rat, endotoxin results in the accumulation of neutrophils in the airspaces of the lung (4,

21). Macrophages recovered from the lungs of endotoxintreated rats spontaneously secrete chemotactic activity for neutrophils (22). We investigated the dose dependency of this response and tested the hypothesis that macrophages from animals treated with endotoxin have an augmented ability to produce chemoattractants. We recovered macrophages from endotoxin-treated rats and then further stimulated these cells by incubating them in the presence of endotoxin. We found that endotoxin is capable of stimulating lung macrophages in vitro to produce chemoattractants. Prior in vivo contact with endotoxin augments the in vitro effect of endotoxin on macrophage production of chemoattractants. This mechanism may be important in perpetuating or intensifying macrophage-mediated neutrophil recruitment in vivo if a similar pattern of enhanced production of chemotactic factors occurs with repeated in vivo stimulation. MATERIALS AND METHODS Animals. Male Sprague-Dawley specific-pathogen-free rats weighing 200 to 280 g (Charles River Canada, Inc., St. Constant, Quebec, Canada) were used exclusively. All animals were studied within 7 days after arrival. These rats received either a "low dose" (2.5 mg/kg) or a "high dose" (5.0 mg/kg) of purified Escherichia coli endotoxin, which was injected into the peritoneal cavity. These doses were found in preliminary experiments to give different sublethal inflammatory reactions. We found in these experiments that the 50% lethal dose for this species was 7.5 mg of our endotoxin preparation per kg. Each experiment was con-

* Corresponding author. t Present address: Center for Lung Research, Vanderbilt University, Nashville, TN 37232.

810

VOL. 57, 1989

ALVEOLAR MACROPHAGE PRODUCTION OF CHEMOATTRACTANTS

trolled by studying a simultaneous cohort of untreated animals. The published experience indicates that the peak influx of neutrophils occurs 24 h after endotoxin injection (4, 21). Thus, we sacrificed the rats at that time. Administration of endotoxin. Lyophilized lipopolysaccharide from E. coli (serotype 055:B5 [Sigma Chemical Co., St. Louis, Mo.]; Westphal phenol extraction method) was stored at 5°C until used. The endotoxin was diluted in sterile phosphate-buffered saline and injected into the peritoneal cavity. Whole lung lavage. After light pentobarbital anesthesia, the chest and abdomen were opened and the rats were exsanguinated by right ventricular puncture. The trachea was exposed and cannulated with a 16-gauge Teflon catheter which was tied into place with 2-0 silk suture. The lungs were then lavaged in situ with 6-ml samples of phosphatebuffered saline until a total return of 45 ml of lavage fluid was collected. The lungs were gently massaged with each infusate. Cell preparation. The lavage fluid was centrifuged at 400 x g at room temperature for 10 min, and the supernatant was decanted and discarded. The cell pellet was suspended in modified (calcium- and magnesium-free) Hanks balanced salt solution. The cells were again centrifuged and washed two more times with modified Hanks balanced salt solution. Total cell count was determined on a grid hemacytometer. Cell viability was monitored by counting the percentage of cells which excluded 0.4% trypan blue dye (GIBCO Laboratories, Chagrin Falls, Ohio). A differential cell count was determined on a cytocentrifuge slide preparation (Cytospin; Shandon Southern Products, Ltd., Runcorn, Cheshire, England) stained with a modified Wright stain (Diff Quick; Dade Diagnostic, Inc., Aquada, P.R.). Four hundred consecutive cells were judged as being AMs or polymorphonuclear leukocytes (PMNs) under oil immersion (x 1,000). Lymphocytes and erythrocytes were not present in significant numbers in the lavage fluid from these rats at any of the time points studied. Production of macrophage-conditioned medium (MCM). Washed lavage cells were suspended in RPMI culture medium (MA Bioproducts, Grand Island, N.Y.) with 10% fetal bovine serum, 20 mM HEPES buffer (N-2-hydroxyethylpiperazine-N'-1-ethane-sulfonic acid), L-glutamine (2.9 mg/ml), penicillin (200,ug/ml), and streptomycin (100 jig/ml). A cell suspension containing 1.2 x 106 viable AMs was plated in a 35-mm culture well (Falcon; Becton Dickinson Labware, Oxnard, Ca.) and incubated at 37°C in 5% CO2 for 60 min. The wells were then vigorously washed three times with Hanks balanced salt solution to remove the nonadherent cells. This method removed 75% of the neutrophils but 90% neutrophils with >90%o viability. Chemotaxis results are expressed as mean values for triplicate measurements of three consecutive column fractions in series. Statistical methods. Differences between levels of in vivo endotoxin exposure with respect to cell recovery and viability were tested by one-way analysis of variance, using Fisher's least significant difference to make pairwise comparisons. The statistical significance of the effects of intraperitoneal endotoxin and in vitro stimulation with endotoxin on the production of chemotactic activity by AM was also assessed by an analysis of variance. Since identical cultures were treated with various concentrations of endotoxin in vitro, a repeated measures design was used. The relationship between the percent PMNs in the lavage fluid and the chemotactic activity of AM from the same lavage was evaluated by regressing the chemotactic index on the natural logarithm of the percent PMN. The relationship between the spontaneous and the stimulated production of chemoattractants was evaluated by linear regression analysis.

RESULTS Animal survival. All control rats (n = 14) and those injected with the 2.5 mg (n = 12) of endotoxin per kg survived to the study point 24 h after treatment. There were two deaths in rats (n = 18) treated with 5.0 mg of endotoxin per kg. All endotoxin-treated rats exhibited tachypnea, diarrhea, and lethargy within 4 to 6 h of injection which improved by the time of sacrifice. Effect of endotoxin on lavage cell recovery. The total number of cells recovered by lavage was not significantly different among the three groups studied. The total cell recovery from the rats that received the high dose of endotoxin was 4.1 + 0.6 x 106 cells compared with 3.4 + 0.3 x 10W and 3.2 ± 0.3 x 106 from rats given the low dose of endotoxin and control rats, respectively. Only rare erthrocytes were seen on the cytocentrifuge preparations, and thus the leukocytes seen were not the result of hemorrhage or extravasation from the vascular compartment. The viability of Javage cells was not different among control rats and rats

INFECT. IMMUN.

A 80-

0ou

o

'1

0

0~~Q

, MCM from rats treated with 5.0 mg of endotoxin per kg in vivo. See text for statistical interpretation.

given 2.5 or 5.0 mg of endotoxin per kg (93 ± 0.8% versus 95.1 ± 0.6% and 94.5 ± 0.7%, respectively). The viability of adherent macrophages was 90% following the 15-h incubation period. This survival in tissue culture was not affected by concentrations of endotoxin in vitro ranging from 0 to 60 ,ug/ml. Only rarely were PMNs seen in the lavage fluid of control rats (1.2 ± 0.5%), but increased numbers were found in animals given 2.5 and 5.0 mg of endotoxin per kg (6.9 ± 2.6% and 26.7 ± 5.8% [P < 0.05 and P < 0.001, respectively]). Thus, endotoxin induced an alveolitis in proportion to dose. Since significant numbers of neutrophils were not recovered from the airspaces of control rats, we think that the animals were free of active respiratory infections. Effect of endotoxin on macrophage production of chemotactic activity. AMs from untreated control rats were stimulated to produce chemotactic activity for PMN when cultured in the presence of endotoxin. There was a direct relationship between the concentration of endotoxin in the culture medium and the chemotactic activity produced by AMs (Fig. 1A).

ALVEOLAR MACROPHAGE PRODUCTION OF CHEMOATTRACTANTS

VOL. 57, 1989 -J 75 T

*200-

0

m :

0

60 ±

En

X_,

0

813

0

45 t

I-

160-

0

120

a

E

0

0

0

1

0

0 0

a) (L) 30 t E

ao

C-) 15t 5)

0,(k

a-

0

0 en

25

50

75

aX 80

0

0

&0

,Q 0 100

125

150

175

40

200

20

Macrophage chemoattractants for neutrophils FIG. 2. Percent neutrophils in bronchoalveolar lavage (BAL) plotted against macrophage production of chemotactic activity (r = 0.857; P < 0.0001). Chemotactic activity produced by the macrophages was stimulated by incubating them in the presence of 30 ,g of endotoxin per ml. Chemotactic activity is expressed as number of neutrophils per 10 oil immersion light-microscopic fields.

AMs from rats injected with the low dose of endotoxin spontaneously produced 50% more chemotactic activity in culture than did control cells. When these AMs were stimulated also in vitro with 15 and 30 p,g of endotoxin per ml during the incubation period, they produced more chemotactic activity than when control AMs were treated with these concentrations of endotoxin. There was no difference in the spontaneous secretion of chemotactic activity by AMs from rats treated with the higher dose (5.0 mg/kg) compared with the lower dose (2.5 mg/kg) of endotoxin, but these high-dose AMs produced augmented amounts of chemotactic activity when stimulated with in vitro exposure to endotoxin (Fig. 1B). Both the in vivo and the in vitro application of endotoxin stimulated the production of chemotactic activity by AMs. Differences between the control group and the rats receiving endotoxin in vivo were significant (P < 0.001) over all levels of in vitro stimulation with endotoxin. Differences in the production of these factors due to in vitro stimulation with endotoxin were also significant (P < 0.001) and were linearly related to dose regardless of the amount of endotoxin given in vivo. The statistical analysis also revealed a significant (P = 0.01) interaction between in vivo and in vitro effects of endotoxin, indicating that the in vitro effect at each dose differs depending on the level of in vivo stimulation. Cells from both the control rats and those receiving 2.5 mg of endotoxin per kg in vivo responded in a qualitatively similar way to in vitro stimulation, but the curves are shifted in magnitude (Fig. 2). For those receiving 5.0 mg of endotoxin per kg in vivo, however, the effect of in vitro stimulation was accentuated and the slope of the curve as well as the magnitude of the response were altered. Since neutrophil contamination could contribute increased chemotactic activity, we analyzed the conditioned medium from peritoneal exudate neutrophils which we incubated in concentrations of 0.4 x 106 to 2.4 x 106 cell per ml. The chemotactic activity of this neutrophil-conditioned medium was not different from that of the negative control. Relationship between lavage parameters and AM production of chemotactic activity. We found a relationship between the percent PMNs in lavage fluid and the unstimulated and endotoxin-stimulated production of chemotactic activity. Lavage specimens with increased PMNs also showed proportionately increased secretion of chemoattractants by AMs. We analyzed these data as a curve, using the formula

40

60

80

100

Spontaneous production of chemotactic activity

FIG. 3. Stimulated production of chemotactic activity by cultured AMs plotted against the spontaneous production of chemotactic activity, both measured as the number of neutrophils per 10 oil immersion light-microscopic fields (r = 0.6422; P < 0.001).

y = a + bln x, where y is the percent PMN in bronchoalveolar lavage and x is the chemotactic activity measured in MCM. There was a significant association between the unstimulated macrophage production of chemotactic activity and the percent PMNs in bronchoalveolar lavage (r = 0.635; P < 0.005). An even stronger correlation between these variables was found when the macrophage production was stimulated by 15 and 30 pLg of endotoxin per ml during the incubation period (r = 0.853, P < 0.001; and r = 0.857, P < 0.001, respectively). These data are summarized in Fig. 2, which plots the relationship between the PMNs in lavage fluid and the stimulated production (with 30 ,ug of endotoxin per ml) of chemotactic activity by AMs. There appears to be a threshold effect in which neutrophil recruitment is evident when the stimulated AM production of chemotactic activity reaches a critical value of 100 neutrophils per 10 oil immersion light-microscopic fields. We noted a significant correlation between the spontaneous production of chemotactic activity and the activity stimulated by 30 jxg of endotoxin per ml (r = 0.6422; P < 0.001) from the AM of each rat (Fig. 3). We interpret this finding as evidence that the spontaneous production of chemotactic activity by AMs predicts the stimulated production of chemotactic activity by the same AMs. Random versus directed migration. The results of a checkerboard experiment, using conditioned medium, are shown in Table 1. These data indicate that the conditioned medium has primarily chemotactic activity for neutrophils with minor chemokinetic properties. Serial dilutions of MCM demonstrated a linear decrease in progressively lost chemotactic activity until control values were reached. As shown, chemotactic activity toward a concentration gradient was markedly increased compared with random migration against a concentration gradient (stimulated random migra-

tion). Partial characterization of AM-generated chemotactic activity. We found no degradation of chemotactic activity when conditioned medium was subjected to heating at 57°C for 30 min, and 87% of the chemotactic activity was retained when the conditioned medium was boiled (100°C) for 30 min. The activity was almost entirely degraded by trypsin exposure, indicating a protein component to its structure. Cycloheximide at a concentration of 10 ,ug/ml inhibited the production of chemotactic activity by 30% but consistently reduced macrophage viability (trypan blue dye exclusion) to 50% of the pretreatment value. Further, when 10 ,ug of

814

INFECT. IMMUN.

CHRISTMAN ET AL.

tractants has not been studied extensively. The mechanism of macrophage activation by endotoxin is unknown, but two

TABLE 1. Random and directed migration of neutrophils stimulated by culture media conditioned by AMs recovered from endotoxin-treated rats' Dilution below filter

0 1:8 1:4 1:2

Chemotactic index at given dilution above filter 1:2 1:8 1:4 0 .0 2.0 .07.7 452. 6 75.5

32

68.0 107.7

2.3 6.5 7

a Values represent the mean number of neutrophils that migrated through polycarbonate filters per 10 oil immersion fields in triplicate chambers (chemotactic index). Values within the diagonals are the responses in uniform concentrations of MCM and thus represent stimulated random migration. Values to the left and right of the diagonals are the responses to positive and negative gradients of MCM, respectively, and thus represent directed migration. The coefficient of variation for the triplicate assays was 27%. b Negative control is Gey balanced salt solution fortified with 1% bovine serum albumin.

cyclohexamide per ml was added to the macrophage-conditioned medium after the incubation period, there was also a 30% reduction in chemotactic potency. Indomethacin in the highest concentration (30 ,xg/ml) tested did not affect cell viability but reduced macrophage production of chemotactic activity by 22%. This small effect is even less signficiant since we observed that, when indomethacin is added following the incubation period, an even greater reduction in chemotactic potency is observed. This finding is consistent with a direct migratory inhibiting effect on neutrophil by indomethacin. There was no chemotactic activity detected in cell lysates (Table 2), although macrophage-conditioned medium contained chemotactic activity. The chemotactic activity of the cell lysate was not increased by prior treatment with the antiprotease phenylmethylsulfonyl fluoride. This cell-associated fraction was not significantly different from the culture medium negative control, while the secreted fraction contained increased chemotactic activity. Figure 4 presents the results of molecular-sieve chromatography performed on a Sephadex G-100 column. A single protein peak was localized by optical density determination at 280 nm just above the termination of the included volume (350 daltons) in fractions 90 through 120. Two peaks of chemotactic activity were measured which represented a 2.5-fold increase over background activity. One peak corresponds to the low-molecular-weight region (near 1,000s) and the second peak is between the 44,000- and the 68,000molecular-weight markers (at approximately 50,000s). DISCUSSION Endotoxin is known to have the capacity to stimulate or activate macrophages (5, 23; Y. Sibille, G. P. Naegel, W. W. Merrell, and H. Y. Reynolds, Am. Rev. Respir. Dis. 131: A388, 1985), but its effect on AM production of chemoat-

major possibilities have been suggested (19). The first claims that there are specific endotoxin receptors on the membranes of sensitive cells. The second hypothesis states that endotoxin provokes nonspecific lipid A-dependent perturbations in the cell membrane. These hypotheses converge to state that this membrane-endotoxin interaction results in a sequence of events which culminates in altered macrophage function. This phenomenon can be exploited in the research laboratory by using endotoxin to stimulate macrophages in vitro to secrete biologic response mediators such as interleukin-1 (3) and macrophage-derived growth factor (12). We have extended this observation by evaluating the effect of endotoxin on AM production of chemotactic factors. Although many factors could result in the accumulation of PMNs in the lung after a pulmonary or extrapulmonary injury, macrophage-generated chemoattractants are believed to be of primary importance (20). In support of this hypothesis, we report a significant correlation between the percentage of neutrophils recovered by lavage and the production of chemotactic activity by AMs from the same specimens. This correlation is consistent with a cause-and-effect relationship in which macrophages produce chemoattractants that result in the accumulation of neutrophils in the lung. AMs from animals with increased percentages of PMNs in their lavage fluid also produce more chemotactic activity when stimulated in vitro, implying an increased ability to produce this monokine. We noted that AMs from endotoxin-treated rats spontaneously produce more chemotactic activity and also have an increased ability to produce chemotactic activity when stimulated in vitro. The amount of chemotactic activity spontaneously produced by the AMs predicts their ability to be stimulated to produce more of this activity. This finding is consistent with a change in the function of resident AMs in response to in vivo provocation. Thus, initial treatment with endotoxin does not maximize or blunt the macrophage response to it, but rather it primes the macrophage for an augmented response upon repeat exposure. In the setting of repeated stimulation, the resident AM could serve as a biologic amplifier of the injurious effect of endotoxin on the lung. The implications this might have for human disease are unclear, since rats are much more resistant to the biologic effects of endotoxin than are humans. The reasons for this difference between the species has not been investigated. The augmentation of AM chemoattractant production by repeated endotoxin stimulation is potentially quite important. Macrophages stimulated in the course of endotoxemia might have an augmented response to ongoing endotoxemia or episodes of repeated exposure. A clinical correlation to our findings is that gram-negative infections are frequently associated with acute lung injury (18). It has been noted that patients who develop endotoxin-associated acute lung injury

TABLE 2. MCM chemotactic activity versus cell lysate chemotactic activity Chemotactic activity (mean ± SEM, n = 3)

Macrophages

Unstimulated Stimulated'

Control

Medium alone

MCM

Cell lysate

27.7 ± 2.1 33.3 ± 1.2

44.3 ± 2.02 110.3 ± 2.4

22.7 ± 2.02 35.3 + 2.4

a Stimulated in tissue culture with 30

p.g of endotoxin per ml.

Endotoxin injection (5.0 mg/kg) MCM Cell lysate

70.3 ± 4.3 113.0 ± 2.5

21.7 ± 4.4 25.7 ± 3.0

ALVEOLAR MACROPHAGE PRODUCTION OF CHEMOATTRACTANTS

VOL. 57, 1989

0.300

-

0.250

-

0.200

+

815

E 0

CO

0.150 +

0 co

0.100

-

0.050 ±

-I-l

0.000

IL.

5

80

-H

-

o

v-

0 s0

+

Q) (a.

404+

0

20-+

S.

Ur)

0 0

VOID

:3

()

z

0

00 (0

L.

U)

0

o-

-4I

l s

40

90

100

A

110

Fraction Number FIG. 4. Optical density at 280 nm (O.D. 280 nm) of fractions 35 to 128 from the G-100 molecular-sieve chromatography column (top) and chemotactic activity of fractions 40 to 110 (average of three consecutive fractions presented as a histogram) (bottom). The molecular weight of the calibration proteins (described in Materials and Methods) are indicated by the arrows. The fraction number is shown on the y axis. Two peaks of chemotactic activity corresponds to a low molecular weight near 1,000 and a higher molecular weight between the 44,000- and 68,000-molecular-weight markers. Chemotactic activity was not identified in the void volume fraction or in the fractions following the termination of the included volume. Chemotactic activity is expressed as the number of neutrophils migrating in 10 oil immersion light-microscopic fields (OIF).

have a particularly poor prognosis (11) and those with acute lung injury complicated with gram-negative infection have an increased incidence of multisystem organ failure (2). Ample opportunity for endotoxin-macrophage interaction to occur exists in hospitalized patients: endotoxins may be produced within the lung during the growth phase of gramnegative pneumonia, they may be inhaled from contaminated respiratory therapy equipment and interact directly with lung cells, or they may circulate to the lung in the course of sepsis from distant sites of infection. Altered permeability of the gastrointestinal tract may result in the absorption of endotoxin from this reservoir as well. Our data implicate an endotoxin-macrophage interaction as a contributor to the inflammatory influx seen in endotoxin-induced alveolitis and further imply that there may be an excessive production of chemoattractants by AM during a repeated or prolonged exposure to endotoxin. This observation is consistent with the data of Snella (24), who reported a trend toward increased chemotactic activity production by AMs from endotoxin-exposed guinea pigs. The physicochemical characteristics of rat AM chemotactic activity have not been reported previously. We measured the chemotactic activity of a factor that is heat stable and degraded by trypsin. The checkerboard analysis (Table 1) suggests that this activity is predominately chemotactic with lesser chemokinetic activity. The secreted and cell lysate experiments suggest that this activity is synthesized just prior to secretion rather than stored in intracellular compartments for later release.

There is activity at a low-molecular-weight region (approximately 1,000) and a high-molecular-weight region (approximately 50,000) of the G-100 column. Extensive but somewhat contradictory information is available regarding this factor in humans and other experimental animals (20). At least two AM-derived factors exist in humans. The best-characterized factor is a low-molecular-weight substance which is probably leukotriene-B4, a lipoxygenase pathway product (16). The low-molecular-weight size of one of the chemotactic activity peaks is consistent with leukotriene-4 or another lipoxygenase pathway product. The destruction by trypsin treatment implies a protein component to the structure of this factor. Although cyclohexamide (a protein synthesis blocker) decreased the production of chemotactic activity by cultured AMs, this could represent a simple cytotoxic effect. Several centers have reported a macrophage- or monocyte-derived chemotactic factor which is protein in nature and has a molecular weight of approximately 10,000 (14, 17, 27). Our factor appears to be closer to 50,000 daltons in size and may represent a trimer or tetramer. This polymerization of biologically active proteins is seen with tumor necrosis factor, which has an apparent molecular weight of 45,000 when separated by gel permeation chromatography and a 17,000-dalton size when purified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (1). Further characterization will be required to define more clearly the molecule(s) responsible for rat macrophagederived chemotactic activity. Other well-described AM-produced monkines may have

816

CHRISTMAN ET AL.

chemotactic activity. We have excluded interleukin-1 as such a factor in our study by documenting the heat stability and the molecular weight of the activity we are measuring, since interleukin-1 has a molecular weight of 17,000 and is degraded by heating to 57°C for 30 min (7). Macrophagederived growth factor for fibroblasts is similar to plateletderived growth factor and may have chemotactic properties for inflammatory cells (6). We have excluded activity due to this factory by our partial characterization, since macrophage-derived growth factor activity is destroyed by boiling (13). In summary, we have shown that AM production of chemotactic activity is increased by endotoxemia and that direct stimulation of AM by endotoxin in vitro results in the increased production of chemoattractants. The combination of in vivo and in vitro endotoxin exposure results in an even greater production of this activity. The ability of the AM to produce chemoattractants is significantly correlated with the intensity of the neutrophilic influx. These data may have clinical implications if endotoxin-macrophage interactions result in AM production of chemoattractants which then cause PMN influx into the lung. This mechanism may partially explain the association between endotoxemia and acute lung injury in humans. Further, recurrent episodes of endotoxemia in the course of acute lung injury could perpetuate and accentuate the injury via a macrophage-mediated neutrophil recruitment mechanism. ACKNOWLEDGMENTS We thank Anne Traynor and David Stump for assistance with the molecular sieve chromatography experiment and for helpful criticism and comments. We also thank Laura Hill-Eubanks for careful technical assistance. This work was supported in part by Public Health Serivce SCOR grant HL-14212 from the National Heart, Lung, and Blood Institute. LITERATURE CITED 1. Aggarwal, B. B., W. J. Kohr, P. E. Hass, B. Moffat, S. A. Spencer, W. J. Henzel, T. S. Bringham, G. E. Nedwin, D. V. Goeddel, and R. N. Harkins. 1985. Human tumor necrosis factor: production, purification, and characterization. J. Biol. Chem. 260:2345-2354. 2. Bell, R. C., J. J. Coalson, J. D. Smith, and W. G. Johanson. 1983. Multiple organ system failure and infection in Adult Respiratory Distress Syndrome. Ann. Intern. Med. 99: 293-302. 3. Bendtzen, K. 1983. Biological properties of interleukins. Allergy 38:219-226. 4. Chang, J. C., and M. Lessor. 1984. Quantification of leukocytes in bronchoalveolar lavage samples from rats after intravascular injection of endotoxin. Am. Rev. Respir. Dis. 125:72-75. 5. Davis, W. B., I. S. Barsoun, P. W. Ramwell, and H. Yeager. 1980. Human alveolar macrophages: effects of endotoxin in vitro. Infect. Immun. 9:753-758. 6. Deuel, T. F., R. M. Senior, J. S. Juang, and G. L. Griffin. 1982. Chemotaxis of monocytes and neutrophils to platelet derived growth factor. J. Clin. Invest. 69:1046-1049. 7. Dinarello, C. A. 1984. Interleukin-1 and the pathogenesis of the acute phase response. N. Engl. J. Med. 311:1413-1419. 8. Eiermann, G. J., B. F. Dickey, and R. S. Thrall. 1983. Polymorphonuclear leukocyte participation in acute oleic acid induced lung injury. Am. Rev. Respir. Dis. 127:845-850. 9. Falk, W., R. H. Goodwin, and F. J. Leonard. 1980. A 48-well microchemotaxis assembly for rapid and accurate measurement

INFECT. IMMUN.

of leukocyte migration. Tissue Cell 17:461-472. 10. Hunninghake, G. W., J. E. Gadek, T. J. Lawley, and R. G. Crystal. 1981. Mechanism of neutrophil accumulation in the lungs of patients with idiopathic pulmonary fibrosis. J. Clin. Invest. 68:259-261. 11. Kaplin, R. L., S. A. Sahn, and T. L. Petty. 1979. Incidence and outcome of the respiratory distress syndrome in gram-negative sepsis. Arch. Intern. Med. 139:867-871. 12. Kovaks, E. J., and J. Kelley. 1985. Secretion of macrophagederived growth factor during acute lung injury induced by bleomycin. J. Leuk. Biol. 37:1-14. 13. Kovaks, E. J., and J. Kelley. 1986. Intra-alveolar release of competence-type growth factor after lung injury. Am. Rev. Respir. Dis. 133:68-72. 14. Kownatzki, E., A. Kapp, and S. Uhrich. 1986. Novel neutrophil chemotactic factor derived from human peripheral blood mononuclear leukocytes. Clin. Exp. Immunol. 64:214-222. 15. Kreisle, R. A., C. W. Parker, G. L. Griffin, R. M. Senior, and J. F. Stenson. 1985. Studies of leukotriene B4-specific binding and function in rat polymorphonuclear leukocytes: absence of a chemotactic response. J. Immunol. 134:3356-3363. 16. Martin, T. R., L. C. Altman, R. K. Albert, and W. R. Henderson. 1984. Leukotriene B4 production by human alveolar macrophages: a potential mechanism for amplifying inflammation in the lung. Am. Rev. Respir. Dis. 129:106-110. 17. Merrell, W. W., G. P. Naegel, R. A. Matthay, and H. Y. Reynolds. 1980. Alveolar macrophage-derived chemotactic factor: kinetics of in vitro production and partial characterization. J. Clin. Invest. 65:268-275. 18. Montgomery, A. B., M. A. Stager, C. J. Carrico, and L. D. Hudson. 1985. Causes of mortality in patients with the adult respiratory distress syndrome. Am. Rev. Respir. Dis. 132: 485-491. 19. Morrison, D. C., and R. J. Uleritch. 1978. The effects of bacterial endotoxin on host mediation systems. Am. J. Pathol. 93:527-617. 20. Reynolds, H. Y. 1983. Lung inflammation: role of endogenous chemotactic factors in attracting polymorphonuclear granulocytes. Am. Rev. Respir. Dis. 127(Suppl.):16-25. 21. Rinaldo, J. E., J. E. Dauber, J. W. Christman, and R. M. Rogers. 1984. Neutrophil alveolitis following endotoxemia. Am. Rev. Respir. Dis. 130:1065-1071. 22. Rinaldo, J. E., J. E. Henson, J. H. Dauber, and P. M. Henson. 1985. Role of alveolar macrophages in endotoxin-induced neutrophils alveolitis in rats. Tissue Cell 17:461-472. 23. Shands, J. W., D. J. Peavy, B. J. Gormus, and J. McGraw. 1974. In vitro and in vivo effects of endotoxin on mouse peritoneal cells. Infect. Immun. 9:106-112. 24. Snella, M.-C. 1986. Production of a neutrophil chemotactic factor by endotoxin stimulated alveolar macrophages in vitro. Br. J. Exp. Pathol. 67:801-807. 25. Snyderman, R., H. S. Shin, and A. M. Dannenberg. 1972. Macrophage proteinase and inflammation. The production of chemotactic activity from the fifth component of complement in macrophage proteinase. J. Immunol. 109:896-902. 26. Tate, R. M., and J. E. Repine. 1983. Neutrophils in the adult respiratory distress syndrome. Am. Rev. Respir. Dis. 125: 552-559. 27. Yoshimura, T., K. Matsushima, J. J. Oppenheim, and E. J. Leonard. 1987. Neutrophil chemotactic factor produced by lipopolysaccharide-stimulated human blood mononuclear leukocytes: partial characterization and separation from interleukin 1. J. Immunol. 139:788-793. 28. Zigmond, S. H., and J. G. Hirsch. 1973. Leukocyte locomotion and chemotaxis. New methods for evaluation and demonstration of a cell derived chemotactic factor. J. Exp. Med. 137: 387-392.