Eur. J. Immunol. 1992. 22: 2687-2695
Anne Boutted, Monique DehouxA, Marianne DeschenesA, Jean-Denis RouzeauO, Phuong Nhi BoriesO and Genevikve DurandO Laboratoire de Biochimie AA, H6pital Bichat, Paris and Laboratoire de Biochimie gCnCraleo, UniversitC Paris XI, Chitenay Malabry
al-Acid glycoprotein potentiates the secretion of IL-IV, IL-6 and TNF-a
2687
al-Acid glycoprotein potentiates lipopolysaccharide-induced secretion of interleukin-1 p, interleukin-6 and tumor necrosis factor-a by human monocytes and alveolar and peritoneal macrophages" Although the physiological role of al-acid glycoprotein (AGP), an acute-phase protein, is poorly understood, several lines of evidence support a modulatory action on the immune response. In this study, we investigated the effect of AGP on the production of interleukin (IL)-@, IL-6 and tumor necrosis factor (TNF)-a by human monocytes, macrophages and the monocytic THP-1 cell line. AGP significantly enhanced (2- to 7-fold) the production of these cytokines in monocytes induced by suboptimal concentrations of lipopolysaccharide [E. coli lipopolysaccharide (LPS): 100 ng/ml] in serum-free conditions, whereas it had little or no effect in the absence of LPS. The potentiating effect of AGP was inhibited by specific antibodies. It was concentration dependent and the greatest enhancement was observed with 250-500 vg/ml. Moreover, AGP only potentiated the effect of suboptimal concentrations of LPS. AGP did not alter the time course of LPS-induced IL-lp, IL-6 or TNF-a secretion. AGP acts as a co-inducer and could also potentiate cytokine secretion triggered by Neisseria meningitidis LPS and muramyl dipeptide. The glycan moiety of AGP did not seem to be involved in its potentiating effect, since both its major glycoforms and asialo-AGP potentiated the effect of LPS to the same extent as native AGP. Possible differences in the effect of AGP according to cell maturation were investigated using isolated human macrophages: AGP potentiated LPS-induced cytokine production by both peritoneal and alveolar macrophages. These data suggest that AGP can modulate monocyte/macrophage functions, thereby contributing to the amplification and regulation of immune and inflammatory responses.
1 Introduction al-Acid glycoprotein (AGP) is one of the best-characterized acute-phase glycoproteins in terms of its peptide sequence and oligosaccharide moiety [l]. Several studies have revealed changes in AGP plasma levels and alterations of the oligosaccharide moiety in response to a number of stimuli. Plagma levels increase by 2- to 4-fold in response to infection and inflammation, and an increase in the proportion of concanavalin A (Con A)-unreactive AGP is observed during pregnancy and liver damage,whereas Con Areactive variants increase during acute inflammatory disorders 111. Although the physiological role of AGP is poorly understood, it may have a modulatory action on the immune response [2].AGP acts on a wide variety of cells involved in the early and late stages of the inflammatory process, including platelets 131, polymorphonuclear cells
[4], lymphocytes (B and T cells) [5-81, and monocyte/ macrophages 19-111. Recent findings concerning the regulation of the immune response suggest that most of these effects of AGP involve cytokines. Some cytokines, especially IL-6, IL-lfl and TNF-a, essentially produced by monocyte/macrophages, are known to regulate the synthesis [l]and glycosylation of AGP by hepatocytes [12]. The action of AGP on monocyte/macrophages, which are involved early in the systemic inflammatory process, has rarely been investigated and the data reported have been controversial [9-l1].This prompted us to study the effect of human AGP and its naturally occurring glycoforms on the secretion of IL-lB, IL-6 and TNF-a by human monocytes and tissue macrophages from the peritoneal and alveolar cavities.
2 Materials and methods 2.1 Reagents [I 101291 @ :
This work was supported by a grant from INSERM (contrat externe 873001).
Correspondence: Monique Dehoux; Laboratoire de Biochimie A, Hopital Bichat, 46 rue Henri Huchard, F-75018 Paris, France Abbreviations: AGP: a - A c i d glycoprotein dipeptide
MDP: Muramyl
0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1992
Escherichia coli (strain 055: B.5)-derived LPS was obtained from Difco Laboratories (Detroit, MI). Neissrria meningitidis-derived LPS was a kind gift from Dr. J. M. Cavaillon (Institut Pasteur, Paris). Muramyl dipeptide (MDP) and phorbol 12-myristate-13-acetate (PMA) were obtained from Sigma (La Verpilliere, France), polymyxin B from Pfizer (Orsay, France). AGP. LPS and MDP were dissolved in RPMI 1640 medium (Gibco, Cergy-Pontoise, France) prior to use.
+
0014-2980/92/1010-2687$3.50 .25/0
2688
Eur. J. Immunol. 1992. 22: 2687-2695
A. Boutten. M. Dehoux, M. Deschenes et al.
2.2 Preparation of AGP
pure according to morphologic criteria and the naphthyl esterase assay (Sigma).
2.2.1 Purification of human AGP and its Con A variants Four different batches of AGP were used in the experiments.Two were obtained from Sigma (St. Louis, MO) and were extensively dialyzed against PBS in sterile conditions before use to eliminate the slight cytotoxicity observed with the crude preparation. Two batches were prepared in our laboratory by means of immunoaffinity chromatography [ 13].Toseparate AGP into Con A-unreactive AGP (peak I) and Con A-reactive AGP (peak II), native AGP was passed through a Con A-Sepharose column using a technique described clsewhere [ 141.These fractions were then loaded on an a-methylmannoside-agarose column (Sigma) to remove traces of Con A. Finally, AGP preparations in PBS were passed through a Detoxigel column (Pierce, Cheshire, GB; 1 x 2 cm), filter sterilized through a 0.2 pm filter (Millex GS filter, Millipore, Molsheim, France) extensively dialyzed and lyophylized. An additive polysulfone filtration was done for some contaminated preparations [15].
2.4 Preparation of human alveolar macrophages Lung macrophages were isolated from the alveolar spaces of healthy male volunteers by means of bronchoalveolar lavage. Briefly, using optic fiber bronchoscopy, one lobe was washed by instilling 5 , 20-ml aliquots of sterile saline
2.2.2 Quality control All preparations of AGP were tested for the presence of LPS in the Limulus amebocyte lysate test (Kabi diagnostica, detection limit 1 pg/ml; Stockholm, Sweden) and preparations containing more than 200 pg/mg protein were discarded. Purity was checked by SDS-PAGE [16] using Coomassie blue, silver staining (Bio-Rad, Watford, GB) and immunoblotting as previously described [lo]. Whatever the preparation tested, purified human AGP was found to migrate as a single band (40-42 kDa) on SDS-PAGE after Coomassie blue staining or after immunoblotting (Fig. 1). No contaminant protein was detected even with silver staining (not shown).
2.3 Preparation of human peripheral blood monocytes (PBM) Peripheral blood from healthy, medication-free volunteers (20 to 30 years of age) was collected onto EDTA. Mononuclear cells were separated by means of Ficoll-Hypaque (Pharmacia Fine Chemicals, Uppsala, Sweden) density sedimentation. Cells were harvested, washed twice and resuspended in RPMI 1640 medium supplemented with Hepes ( 5 mM), penicillin (100 mU/ml), streptomycin (100 pg/ml) and L-glutamine (2 mM) (complete medium). The complete medium was shown to be endotoxin-free in the Limulus amebocyte lysate assay. The cell suspension was adjusted to a density of 5 x lo6 cells/ml. All experiments were performed with complete medium free of FCS to avoid its potentiating effect [17, 181 and the addition of exogeneous AGP. The monocyte population was purified by 2 h of adherence in 24-well plates (Falcon, Oxnard, CA) at 37 "C, with 5% CO2. At the end of the incubation period, adherent cells were washed three times with HBSS to remove nonadherent cells and then incubated in fresh complete medium.The cell population obtained with this method is 95% to 99%
1
1
3
2
2
3
4
4
Figure I . (A) Coomassie blue-stained SDS-PAGE of human AGP from Sigma (lane 1) and AGP prepared in our laboratory (lanes 3-4). Lane 2 is a molecular mass marker. (B) Immunoblotting of human AGP from Sigma (lanes 1-2) and AGP prepared in our laboratory (lanes 3-4). Samples were analyzed by SDS-PAGE, transferred to nitrocellulose membranes and immunoblotted with specific antibodies against AGP.
Eur. J. Tmmunol. 1992. 22: 2687-2695
al-Acid glycoprotein potentiates the secretion of IL-lp, IL-6 and TNF-tw
and aspirating after each aliquot. The cells were removed from the lavage fluid by means of centrifugation (600 x g, 10 min). Bronchoalveolar fluid contained more than 95% of alveolar macrophages, with a few lymphocytes. The cell suspensions were adjusted to lo6 cells/ml in complete medium. The alveolar macrophage population was purified by 2 h of adherence to plastic dishes. Pulmonary mononuclear phagocytes were more than 95% pure, as determined by nonspccific esterase staining.
2.5 Preparation of human peritoneal macrophages Peritoneal macrophages were obtained by means of peritoneal fluid aspiration from women undergoing laparoscopy for tuba1 ligation. Oral informed consent on the use of the samples for research was obtained from each patient. Cells were separated on Ficoll Hypaque, adjusted to lo6 cells/ml in complete medium and then prepared as described above for alveolar macrophages. Preparations routinely contained more than 95% of macrophages, as determined by nonspecific esterase staining.
2.6 Culture of THP-1 cells THP-I cells were grown in complete medium supplemented with 10% heat-inactivated FCS. THP-1 cells were prepared for experiments by addition of PMA (30 ng/ml), then plated in 24-well dishes (2 x lo5cells in 1 ml) and allowed to differentiate for 48 h [19]. Cell monolayers were washed twice with HBSS prior to the addition of the stimuli. 2.7 Cell stimulation Monocytes, THP-1 cells and macrophages were incubated in the presence or absence of E. coli LPS at several concentrations, with or without various amounts of AGP for various periods of time at 37 “C in a 5% CO2 - air atmosphere. N . meningitidis LPS or MDP were used in some experiments in the presence of polymyxin B. Control cultures included cells cultured alone and in the presence of LPS or MDP alone. In some experiments, AGP was preincubated overnight at 4 “C and continuous shaking with purified specific anti-AGP immunoglobulins [20], prior to addition in cell cultures. After the incubation periods, the supernatants were collected, centrifuged and stored at - 70 “C until assay.
triplicate cultures. A t least three different PBM preparations were used for each experiment and gave similar patterns of results. A “representative experiment” is a typical result obtained with cells from a single donor. Cytokine concentrations were compared using Student’s paired t-test.
3 Results 3.1 AGP alone does not induce significant IL-16, IL-6 or TNF-a secretion by human monocytes Unstimulated PBM produced low levels of IL-1(3,IL-6 and TNF-a. PBM incubated with AGP (250 yg/ml) produced similar or higher levels of cytokines than unstimulated controls, without overstepping a 2- to 6-fold increase (Fig. 2). However, the mean difference between the stimulated and unstimulated cells was not statistically significant regardless of the donor or the AGP preparation used. Higher levels of AGP (500 and 1000 pg/ml) did not yield different results (data not shown), although PBM from the same donors showed marked dose-dependent cytokine production on stimulation by LPS.The PMA-differentiated monocytic THP-1 cell line, which produces cytokines in response to LPS, was unresponsive to AGP alone (not shown). Thus AGP by itself is a poor inducer of cytokine production.
3.2 AGP potentiates LPS-induced secretion of IL-1P, IL-6 and TNF-a by human monocytes AS shown in Fig. 3, AGP (250 pg/ml) significantly enhanced the production of IL-1P (p = 0.026), IL-6 0, = 0.001) and TNF-a (p = 0.001) by human PBM ( n = 11) induced by E. coli LPS (100 ng/ml). I n a more reproducible system, AGP enhanced cytokine secretion by PMA-primed THP-1 cells in response to LPS by 2- to 3-fold (Fig. 3 ) . To determine the specificity of the potentiating effect of AGE AGP was preincubated with purified, specific immunoglobulins before the addition of LPS toTHP-l cell cultures. As shown in Table 1, this preincubation had no effect on cytokine secretion induced by AGP alone. By contrast, the Table 1. The potentiating effect of AGP is suppressed by specific antibodies against AGPa)
IL-lfi
2.8 Enzyme immunoassays for IL-lP, IL-6 and TNF-a The amount of immunoreactive IL-lP, IL-6 and TNF-a in the cell supernatants was determined using enzyme immunoassays (EASIA Medgenix, Fleures, Belgium). The sensitivity of the ELISA tests for each cytokine was 10 pg/ml. Results are expressed as the mean of duplicate assays. Standard errors were always below 10%.
2.9 Statistical analysis Triplicatc monocyte cultures were used for each stimulus concentration. All measurements were done on pooled
2689
Control AGP Ab AGP + Ab LPS LPS + AGP LPS + AGP
+
IL-6
TNF-a
174 f 50 32 k 6 148 f 81 202 f 128 30 k 8 216 f 102 151 f 32 26k 5 140 f 40 208 f 123 29 k 5 198 f 80 2009 f 241 16001?:100 2167 f 600 4685 f 507 3500 f 500 8700 f 2400 Ab 2034 f 842 1575 f 106 2500 f 800
a) THP-1 cells were stimulated or not with E. coli LPS (100 nglml) in the presence or absence of AGP (250 pg/ml) and specific antibodies against AGP (Ab: dilution 1/100). The levels of cytokines (pg/ml) were assayed in the supernatants of 24 h cultures. Results are expressed as mean & SD of two different experiments.
Eur. J. Immunol. 1992. 22: 2687-269s
A. Boutten, M. Dehoux, M. Deschenes et al.
2690
addition of anti-AGP immunoglobulins completely inhibited the potentiating effect of AGE Taken together, these results argue that AGP is directly responsible for the potentiating effect and exclude a role of endotoxin contaminants.
of production obtained with the optimal concentration of LPS (Fig. 4). In the presence of AGP, the levels of immunoreactive cytokines rose, depending on the donor, between 2- and 7-fold relative to those induced by suboptimal doses of LPS alone. Similar results were obtained using a biological assay for IL-1, IL-6 and TNF (data not shown).
3.3 AGP potentiates suboptimal levels of LPS In most experiments a concentration of 100 ng/ml E. coli LPS induced suboptimal cytokine production in the absence of FCS. Indeed, cytokine secretion was E. coliLPS concentration dependent (0.1-5000 ng/ml) and was potentiated by AGP (250 pg/ml) up to, but not beyond the level
3.4 The effect of AGP is concentration dependent AGP showed a concentration-dependent potentiation of IL-lfi, IL-6 and TNF-a production induced by 100 ng/ml
I L - 1 I3
-E
.
cn
n
2500
c? 7-
d
0
PBM
'
10
LPS
RPMl
IL-6
AGP
IL-6
10000
p=O.001 n=l1
-
.
100000
T
E
cn
-5m
THP-1
n v
1WOO
5000
?
i
1000
a
v
4
"-
loo
0
PBM
lo
THP-1
TNF-o(
1
20000 p=0.001 n=ll
-
TNF-S
.
T
E
cn
T
n
B
v
d
i
loooo
z l-
E
a
3 0
I]
PBM I]
1 '
LPS
APMl
AGP
Figure 2. Effect of AGP on 1L 1B. IL-6 and TNF-a synthesis. PBM from 10 different donors were incubated with various concentrations of AGP or E. coli LPS (100 ng/ml). Cytokines levels were assayed in thc supernatants of 24-h cultures. The horizontal lines denote mean valucs.
THP-1
Figure 3. Potentiating activity of AGP on LPS-induced IL-lP, IL-6 and TNF-a secretion by human monocytes from peripheral blood (PBM) and PMA-treated THP-1 cells. Cells were stimulated with E. coli LPS (100 ng/ml) in the presence (El) or absence (m)of native AGP (250 pg/ml). Cytokine levels were assayed in the supernatants of 24-h cultures. Values are mean f SD of experiments with PBM from 11 different donors and of 4 experiments withTHP-1 cells. Significance was evaluated using Student's paired t-test .
Eur. J. Immunol. 1002. 22: 2687-2695
ul-Acid glycoprotein potentiates the secretion of IL-lp, IL-6 and TNF-u
2691
E. coli LPS (Fig. 5). IL-lp, IL-6 and TNF-a synthesis were 3.5 AGP does not alter the time course of IL-lp, IL-6 increased with AGP concentrations as low as 25 or or TNF-a secretion 50 &ml, depending on the donor. The intensity of the potentiating effect of AGP varied greatly according to the The kinetics of LPS-induced secretion of IL-113, IL-6 and donor and the optimal effect for each cytokine was obtained TNF-a in the absence of AGP and FCS were in agreement with concentrations ranging from 250 to 500 pg/ml, i.e. with the results of previous studies [21]. AGP (250 pg/ml) within the range of physiological plasma levels. Cytokines potentiated the secretion of each cytokine but the time levels were steady or tended to fall at concentrations of 1000 course was unchanged, despite some variations in total and 2000 kg/ml (data not shown). cytokine production from experiment to experiment (Fig. 6). In some experiments, AGP strongly potentiated IL-6 secretion, whereas this was not the case for IL-113 and TNF-a, possibly because IL-113 and TNF-a synthesis can be IL-10 inhibited by IL-6 [22]. This led us to choose an incubation 6000 period of 24 h, a time during which cytokine interaction is minimal
r
I
. E
-
4000
IL-1B 8000
e I-
1 -
E . -a
2000
CI
0
0
0.1
1
10
LPS
4000
-
1 0 0 1000 5000
c2
(nglrnl)
2
0
0
I L-6 30000
5 0 100
200
300
AGP
(pg/ml)
400
500
400
500
400
500
1
--
.
IL-6
r
D
n
I
(0
0
0.1
I
10
100
LPS
1000 5000
0
(ng/rnl)
0
5 0 100
230
300
AGP
(pg/ml)
TNF-v TNF-a
I
0
0.1
?
10
LPS
I
100 1 0 0 0 5 0 0 0
(ng/rnl)
Figure, 4. LPS concentration dependence of the potcntating effect of AGP on LPS-induced secretion of IL-113, IL-6 and TNF-a by PBM. PBM were stimulated with increasing amounts of E. coli LPS (0.1-5000 ng/ml) in the presence ( 0 )or absence (0)of native AGP (250 pg/ml). One representative experiment from four.
0
50 1 0 0
223
AGP
300
(lg/ml)
Figure 5. Concentration-dependent effect of AGP on LPSinduced secretion of IL-lP, IL-6 and TNF-a by PBM. PBM were stimulated with increasing amounts of AGP (25-500 kg/ml) in the presence ( 0 )or absence (0)of E. coli LPS (100 ng/ml) for 24 h. Cytokines were assayed in the supernatants of 24-h cultures. One representative experiment from four.
Eur. J. Immunol. 1992. 22: 2687-2695
A. Boutten. M. Dehoux, M. Deschenes ct al.
2602
3.6 Con A-unreactive, Con A-reactive AGP and asialo-AGP have the same potentiating effect as native AGP
I L - 1B
The microheterogeneity of the glycan moiety of AGP is evidenced by its reactivity with the lectin Con A , which can be used to separate it into two major variants: Con A-unreactive and Con A-reactive AGP. The immunosuppressive capacity has been reported to reside in the carbohydrate moiety of AGP [6, 10, 111. Moreover, the binding of AGP to macrophage is closely dependent on the glycan moiety. Thus, it seemed interesting to compare the effects of Con A-reactive, Con A-unreactive AGP and native AGP on LPS-induced cytokine secretion by human monocytes. All three AGP variants significantly increased the LPS-induced production of immunoreactive IL-lD, IL-6 and TNF-a and the potentiating effect of Con A-reactive and Con A-unreactive variants did not differ significantly from that of native AGP (Table 2). Moreover, their effect was not abolished by desialylation (data not shown).
0 8
4
0
‘6
24
32
40
48
HOURS ATER STIMULATION
IL-6 MOoo
1
3.7 Comparative effect of AGP on human PBM and human macrophages 0
4
8
12;
24
32
40
48
Macrophages differ from PBM in their ability to produce IL-lp, IL-6 and TNF-a [23]. To investigate a possible difference in the effects of AGP according to cell maturation, we investigated the action of AGP on human macrophages. AGP potentiated IL-lP, IL-6 and TNF-a secretion of human alveolar and peritoneal macrophages induced by 100 ng/ml E. coli LPS in the absence of FCS by two to three times (Table 3 ) .
HOURS AFTER STIMULATION
TNF-o(
3.8 The potentiating effect of AGP is not specific to E . Coli LPS 3
3
16
21
32
40
48
HOURS AFTER STIMULATION
Figure 6 . Time course of IL-10, IL-6 and TNF-a production by LPS-stimulated monocytes (100 ng/ml), with or without 250 Fglml native AGP Cytokines were assayed in the supernatants collected at the times indicated and kept in the incubator until the end of the cxperirnent. Monocytes incubated with RF’MI (A), AGP (0)LPS (A)and AGP + LPS ( 0 ) .One rcpresentative experiment from four.
To define further the role of AGP in cytokine regulation, we determined whether AGP also influenced cytokine secretion induced by MDP and N. meningitidis LPS. Other experiments were performed with polymyxin B to neutralize possible endotoxin in the AGP preparations, since polymyxin B inhibits various biological activities of LPS both in vivo [24] and in v i m [25]. The inhibition of LPS-induced cytokine secretion by polymyxin B depends on the origin of the LPS. Cavaillon et al. showed that polymyxin B (2 pg/ml) completely inhibited cytokine secre-
Table 2. Con A-unreactive and Con A-reactive AGP have the same potentiating effect as native AGPa)
LPS
IL-lP
IL-6
TNF-a
Control
-
-
Con A-unreactive AGP Con A-reactive AGP
-
31 f 13 2672 k 822 60 f 34 4719 f 390 44 f 21 3607 f 874 54 f 26 4327 f 787
250 f 100 1695 f 343
Native AGP
55 f 16 2636 f 969 148 f 65 8562 f 2303 112 k 70 12940 f 5501
+ +
+ -
+
121 f 68 11OOo f 3937
797 f 402 4025 f 746 447 k 99 4225 f 867 653 f 139 4400 f 697
a) PBM were stimulated (+) or not (-) with E. coli LPS (100 ng/ml) in the presence or absence of native, Con A-reactive, or Con A-unreactive AGP (250 pg/ml). Cytokines (pg/ml) were assayed in the supernatants of 24 h cultures. Results are expressed as mean k SD of four different donors.
Eur. J. Immunol. 1992. 22: 2687-2695
al-Acid glycoprotein potentiates the secretion of IL-lfi, IL-6 and TNF-u
tion induced by up to 2 yg/ml E. coli LPS and reduced the activity of several endotoxins derived from gram-negative bacteria, whereas it only inhibits very low concentrations of N . merzingitidis LPS ( 5 1 ng/ml) [25]. As shown inTable 4, polymyxin B (2 &ml) abolished cytokine secretion induced by E. coli LPS in the presence and absence of AGl?
2693
AGP significantly enhanced cytokine secretion induced by N . meningitidis LPS (10ng/ml) even in the presence of polymyxin B. MDP-induced cytokine secretion was also potentiated by AGP in the presence of polymyxin B to the same extent as that induced by LPS and AGP (Table 5 ) . These results are not in favor of endotoxin contamination and again show that the potentiating effect of AGP is not restricted to LPS.
Table 3. Effect of AGP on LPS-induced secretion of IL-lfl, IL-6 and TNF-a by human peritoneal and alveolar macrophagesa)
4 Discussion
Control AGP LPS LPS + AGP
IL-lg
PM IL-6 TNF-a IL-lP
AM IL-6 TNF-a
29 331 1912 6142
100 312 5320 7420
481 522 1127 1248 1875 17950 4320 24190
64
28 78 570 1180
1398 3370 5380
a) Peritoneal (PM) and alveolar macrophages (AM) were stimulated with E. coli LPS (100 ng/ml) in the presence or absence of native AGP (250 yglml). The levels of cytokines (pg/ml) were assayed in the supernatants of 24-h cultures. One representative experiment from three.
Table 4. The potentiating effect of AGP is not restricted to E. coli LPS")
Control AGP AGP + PB E. coli LPS E. coli LPS AGP E. coli LPS + PB E. coli LPS PB AGP N. m. LPS N. m. LPS PB W. in. LPS + PB + AGP
+ + +
+
Llf3
IL-6
40 250 320 1200 2770 40
37 65 60 ND ND ND
4 0 N D 4000 6000 10800
1000 1095 15465
TNF-a 160
400 440
3000 19000 440
1600 7 800 19000 26 OOO
a PBM were stimulated with E. coli LPS (100ng/ml) or N . rneningitidis LPS (N. rn. LPS, 10ng/ml) in the presence or absence of AGP (250 pg/ml) and polymyxin B (PB: 2 pg/ml). Cytokines (pg/ml) were assayed in the supernatants of 24-h cultures. One representative experiment from three.
Table 5. Effect of AGP on MDP-induced secretion of IL-lP, IL-6 and TNF-a")
IL-lP
IL-6
TNF-a
Control AGP MDP MDP
133 f 60 7 4 f 60 88k29 195 f 85 103 f 76 92 f 30 + PB 935 k 210 865 k 162 940 k 367 + PB + AGP 5340 f 3140 3425 k 2015 3750 k 2616
+ PB
a) PBM were stimulated with MDP (10 pg/ml) and polymyxin B (PB: 2 pg/ml) in the presence or absence of AGP (250 yg/ml). The levels of cytokines (pg/ml) were assayedin the supernatants of 24-h cultures. Results are expressed as mean It SD of three different experiments.
Our findings indicate that human AGP significantly enhances LPS-induced secretion of IL-1fi, IL-6 and TNF-a in serum-free cultures of human monocytes. Both serum and plasma enhance LPS-induced secretion of cytokines [18, 251 and a wide variety of substances isolated from serum have recently been reported to enhance cytokine secretion by human monocytes, including C3a [26], C5a [27], opioids [28], substance P [29], a-adrenergic agonists [30],PAF [31], IL-3 [32], LPS-binding protein [33] and lactoferrin [34]. However, synergism between LPS and any of the above substances could not alone explain the effect of the serum. AGE active at concentrations within the physiological serum range, could thus be one of the potentiating components of the serum. The mechanism by which serum enhances LPS-induced cytokine secretion is poorly understood. C5a [27] and IL-3 [32] activate the transcription of cytokine genes, with or without mRNA translation. Other factors could interact with the LPS receptor; neither the complexity of the LPS receptor on monocytes nor the interaction between LPS and its receptor and the subsequent alterations in the cell have been entirely elucidated [35, 361. Recent reports indicate that several proteins present in human serum, e.g. LPS-binding protein [33] and lactoferrin [34], bind to LPS and can interfere with its effects. Several findings suggested that the potentiating effect of AGP did not involve contaminating LPS. First, it was concentration dependent, with a minimal effective concentration of 25 or 50 pg/ml AGP depending on the donor and an optimal concentration of 250 pg/ml. Higher concentrations consistently induced similar or slighlly decreased cytokine secretion. Second, the effect of AGP was caused by low levels of contaminating LPS, one would expect an additive rather than a synergistic effect. Third, the potentiating effect of AGP was completely suppressed by preincubation of AGP overnight with specific immunoglobulin and by heat denaturation (data not shown). These experiments clearly show the specific role of AGP and exclude the participation of endotoxin contaminants.
In some experiments, moderate levels of cytokines were obtained with AGP alone. However, in additional experiments, when AGP was submitted to filtration on polymyxin B columns and ultrafiltration on polysulfone [15], cytokine secretion was as low as that in controls. After a pretreatment of AGP by heating or with anti-AGP immunoglobulins, the cytokine production observed with AGP alone was unaffected. Thus, AGP alone seemed to be a poor inducer of cytokine secretion. By contrast, LPS-induced cytokine secretion was 2- to 6-fold increased in the presence of AGP. The synergism
2694
Eur. J. Immunol. 1992. 22: 2687-2695
A. Boutten, M. Dehoux, M. Deschenes et al.
induced by AGP showed marked variations according to cell donor. Inter-individual differences in cytokine secretion have been shown to be regulated by class11 MHC molecules [37], but we found no relationship between high cytokine secretors and a high response to AGP A reproducible potentiating effect of AGP was observed when THP-1 cells were used.
human peritoneal and alveolar macrophages. As with PBM, we found no IL-1-inhibiting activity, but rather a synergism between LPS and AGP with both peritoneal and alveolar macrophages.
Conflicting studies have found both immunostimulatory [8] and immunosuppressive [9] effects of AGP on human mononuclear cells. Our results contrast with the 30% maximal inhibition of TNF-a secretion reported by Scuderi et al. [9]. Similarly, although Singh et al. [S] reported a direct and concentration-dependent inducing effect of AGP on IL-1 secretion, we found similar results with AGP preparations containing more than 200 pg/mg contaminating endotoxin. After carefully discarding the contaminated AGP, we did not report any direct effect of AGP but actually a potentiation of LPS-induced cytokine secretion.
We thank Dr. .I. M . Cavaillon of the Institut Pasteur, Paris, for helpful discussions.
AGP is highly glycosylated (40%) and sialylated (1lY0). The glycan moiety has been implicated in the immunosuppressive effects of AGP [6, 10, 111. Moreover, the monoAGP only potentiated the effect of suboptimal concentra- cytes/macrophages possess membrane receptors for both tions of LPS, without overstepping the effect of optimal sialyl [40] and mannose/fucose residues [41]. Con Aconcentrations (5 pg/ml E. coli LPS). In addition, AGP did reactive AGP has been reported to bind to macrophage not alter the time course of LPS-induced IL-lP, IL-6 or lectin. Binding inhibition by mannosylated bovine serum TNF-a secretion and was synergistic with LPS from two albumin suggests the participation of the mannose receptor bacterial species. Polymyxin B inhibits cytokine secretion [42]. We found that native AGE its two major glycoforms induced by E. coli LPS but not by N. meningitidis LPS in (Con A-unreactive and Con A-reactive) and asialo-AGP vivo [24] and in vitro [25]. Complexing E. coli LPS with potentiated the effect of LPS to the same degree, suggestpolymyxin B completely inhibited cytokine secretion ing that the binding of AGP via its glycan moiety to induced by the association of AGP and E. coli LPS, mannose/fucose or sialic acid receptors does not play a part whereas AGP and N. meningitidis LPS were still synergis- in the described phenomena. tic. The potentiating effect of AGP is not restricted to LPS since MDP-induced cytokine secretion was also enhanced The plasma a-globulins are well suited to fill a role as molecules of the immune response. Their secretion by by AGF! hepatocytes can increase rapidly in response to the assoIn preliminary experiments,we found that a 2-h exposure of ciation of IL-1, IL-6 and TNF-a. During local inflammation monocytes to AGP (250 pg/ml) before the removal of AGP and systemic infection, AGP could contribute to enhancing and the addition of E. coli LPS (100 ng/ml) potentiated cytokine production by monocytes and macrophages. The cytokine secretion to an extent comparable to that resulting release of IL-1, IL-6 and TNF-a could trigger and obtained by incubation with both LPS and AGP for 24 h. maintain elevated acute-phase protein synthesis in the liver. These results suggest a priming effect of AGP, as has been Consequently, plasma AGP may function as an element in described with C5a [27].Whether AGP acts at the level of the complex regulatory network that controls the inflamgene expression or the mRNA remains to be determined. matory process.
Previous studies by our group have found a factor inhibiting the activity of IL-1, secreted by murine peritoneal macrophages exposed to human AGP [lo]. However, we have been unable to detect a specific IL-1-inhibiting factor in the culture supernatants of PBM exposed to AGP The comitogenic thymocyte assay used to determine IL-1 activity in the previous study can be influenced by other cytokines and inhibitors [38]. In our study, we found similar results in ELISA and in the biological assays [thymocyte assay, L929 cytotoxic assay and B9 assay for IL-1, TNF and IL-6, respectively (data not shown)]. The discrepancy between the results observed with murine macrophages and human monocytes could be related to at least two factors, i.e. the difference between the activation state of the cells studied and the species specificity of human AGF! Differences between human and murine AGP have been described with respect to their amino acid sequence, glycosylation and inflammatory responses [39]. To investigate possible differences in the effect of AGP according to cell maturation, we further studied isolated
Received November 19, 1991; in final revised form July 8, 1992.
5 References 1 Durand, G., Prog. Clin. B i d . Res. 1989. 300: 241. 2 Krerner, J. M. H., Wilting, J. and J. L H. M., fharmacol. Rev. 1988. 40: 1. 3 Costello, M., Fiedel, B. A. and Gewurz, H., Nuture 1979.281: 677. 4 Laine, E., Couderc, R., Roch-Arveiller, M., Vasson, M. l?, Giroud, J. P. and Raichvarg, D., Inflammation 1991. 14: 1. 5 Chiu, K., Mortensen, R., Osmand, A. and Gewurz, M., Immunology 1977. 32: 997. 6 Bennett, M. and Schmid, K., Proc. Natl. Acad. Sci. USA 1980. 77: 6109. 7 Cheresh, D., Haynes, D. and Distasio, J., Immunology 1984. 51: 541. 8 Singh,V. K., Prog. Clin. Biol. Res. 1989. 300: 283. 9 Scuderi, P., Dorr, R. T., Liddil, J. D., Finley, P. R., Meltzer, l?, Raitano, A. B. and Rybski, J., Eur. J. Immunol. 1989.19: 939. 10 Bories, l? N., Feger, J., Benbernou, N., Rouzeau, J. D., Agneray, J. and Durand, G., Inflammation 1990. 14: 315. 11 Pos, O., Oostendorp, R. A. J. ,Van der Stelt, M. E., Schcper, R. J. and Van Dijk, W., Inflammation 1990. 14: 133. 12 Pos, O., Aarden, L. ,Yap, S. H. and van Dijk,W., in Abstracts of the 7th International Congress of Immunology-Berlin, Gustav Fischer, Stuttgart 1989, p. 450. 13 Biou, D., Konan, D.; Feger. J., Agneray. J., Durand, G.. Leroy, Y , Cardon, l?, Fournet, B., Biochim. Biophys. Acta 1987. 913: 308. 14 Bayard, B., Kerckaert, J. P., Biochem. Biophys. Res. Commun. 1980. 95: 771.
Eur. J. Immunol. 1902. 22: 2687-2695
al-Acid glycoprotein potentiates the secretion of IL-lP, IL-6 and TNF-a
15 Schindler. R., Dinarello, C. A., J. Immunol. Methods 1989. 116: 159. 16 Laemmli, U. K., Nature 1970. 227: 680-685. 17 Cavaillon. J. M. and Haeffner-Cavaillon, N., Zmmunol. Lett. 1985. 10: 35. 18 Lonnemann. G., Endres, S.,Van der Meer, J. W. M., Cannon, J. G.. Koch. K. M., Dinarello, C. A., Eur. J. Zmmunol. 1989. 19: 1531. 19 Golenbock. D.T., Hampton, R.Y., Qureshi, N.,Takayama, K., Ractz, C. R. H.. J. Riol. Chern. 1991. 266: 19490. 20 Tissot, B., Scta, N . , Durand, G., Feger. J., Clin. Chem. 1990. 36: 666. 21 Tc Vclde. A. A., Huijbcns, R. J. F., Heije, K., de Vries, J. E., Figdor. C. G., Blood 1990. 76: 1392. 22 Schindler, R., Mancilla, J., Endres, S., Ghorbani, R., Clark, S. C. and Dinarello, C. A , , Blood 1990. 75: 40. 23 Wewcrs, M. D. and Hcrzyk, D. J., J. Immunol. 1989.143: 1635. 24 Baldwin, G., Alpert, G.. Caputo, G. L., Baskin, M., Parsonnet, J.. Gillis. Z . A.,Thompson. C., Siber, G. R., Fleisher, G. R., J. I n f d Dis. 1991. 164: 542. 25 Cavaillon, J. M., Haeffner-Cavaillon, N., Mol. Immunol. 1986. 23: 965. 26 Cavaillon, N . H., Cavaillon, J. M., Laude, M. and Kazatchkine. M. D., J. Immunol. 1987. 139: 794. 27 Schindlcr, R.. Gelfand, J. A , , Dinarello, C. A., Blood 1Y90. 76: 1631.
28 A p e . R. N., Durum, S. K. and Oppenheim, J., Immunol. Lett. 1990. 24: 141. 29 Lotz. M. ,Vaughan, J. G. and Carson, D. A.. Science 1988.241: 1218.
2695
30 Spengler, R. N., Allen, R. M., Remick, D. G., Strieter, R. M. and Kunkel. S. L., J. Immunol. 1990. 145: 1430. 31 Poubelle, P. E., Gingras, D., Demers, C., Dubois, C., Harbour, D., Grassi, J. and Rola-Pleszczynski, M., Zmmunology 1991. 72: 181. 32 Frendl, G., Fenton, M. J., Beller, D. I., J. Immunol. 1990.144: 3400. 33 Schumann, R. R., Leong, S. R., Flaggs, G. W., Gray, P. W., Wright, S. D., Mathison, J. C., Tobias, P. S., Ulevitch, R. J., Science 1989. 249: 1429. 34 Miyazawa, K., Mantel, C., Lu, L., Morrison, D. C. and Broxmeyer, H. E., J. Immunol. 1991. 146: 723. 35 Hara-Kuge, S., Amano, F., Nishijima, M. and Akamatsu,Y, J. Biol. Chem. 1990. 265: 6606. 36 Shinomiya, H., Hirata, H. and Masayasu, N., J. Immunol. 1991. 146: 3617. 37 Pere, Santamaria, Gehrz, R. C., Bryan, M. K., Barbosa, J. J., J. Immunol. 1989. 143: 913. 38 Dinarello, C. A., Blood 1991. 77: 1627. 39 Cooper, R. and Papaconstantinou, J.. J. Biol. Chem. 1986.261: 1849. 40 Gessl, A., Boltz-Nitulescu, G., Wiltschke, C., Holzinger, C., Nemet, H., Pernerstorfer,T. and Forster, O., J. Immunol. 1990. 142: 4372. 41 Stahl, P. D., Rodman, J. S., Miller, M. J., Schlesinger, P. H., Proc. Natl. Acad. Sci. USA 1978. 75: 1399. 42 Pimpaneau, V , Midoux, P., Durand, G., de Baetselier, P., Monsigny, M. and Roche, A. C., Glycoconjugate J. 1989. 6: 561.