Published online before print February 24, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Infectious Diseases, Faculty of Medicine, Imperial College, Hammersmith Campus, London, United Kingdom
1 Correspondence: Department of Infectious Diseases, Imperial College, Hammersmith Hospital, Du Cane Road, London, W12 0NN, UK. E-mail: j.friedland{at}imperial.ac.uk
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
(TNF-) and interleukin (IL)-1, key cytokines in granuloma formation, in networks involving human osteoblasts and monocytes. Experiments focused on CXC ligand (CXCL)8, CCL2, and matrix metalloproteinase (MMP)-9, human monocyte-derived mediators involved in control of leukocyte influx. TNF- but not IL-1 has a key role stimulating CXCL8 secretion in Mycobacterium tuberculosis-infected human osteoblast MG-63 cells. Conditioned medium from M. tuberculosis-infected osteoblasts (COBTB) drives CXCL8 and some CCL2 gene expression and secretion from primary human monocytes. IL-1 receptor antagonist and to a lesser extent anti-TNF- inhibited COBTB-induced CXCL8 secretion (P<0.01) but did not affect gene expression. IL-1 blockade had a comparatively lesser effect on CCL2 secretion, whereas anti-TNF decreased CCL2 concentrations from 7840 ± 140 to 360 ± 80 pg/ml/4 x 105 cells. Neither proinflammatory mediator affects MMP-9 secretion from COBTB-stimulated human monocytes. In summary, in a paracrine network, M. tuberculosis-infected osteoblasts drive high-level CXCL8, comparatively less CCL2, but do not alter MMP-9 secretion from uninfected human monocytes. This network is, in part, regulated by IL-1 and TNF-.
Key Words: CXCL8 • CCL2 • TNF • IL-1 • networks • gelatinase B
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Chemokines are central in recruitment of specific leukocyte populations during infection and inflammation [4 , 5 ]. CXC ligand (CXCL)8 [or interleukin-8 (IL-8)], the best-characterized C-X-C chemokine, is chemotactic for neutrophils, T lymphocytes, and monocytes [6 7 8 ]. In contrast, the C-C chemokines, of which CCL2 (monocyte chemoattractant protein-1) is a prototypic example, selectively recruit mononuclear cells. Chemokines have been demonstrated in vivo and in vitro to be important during tuberculous infection [9 10 11 12 13 ]. Osteoblast chemokine secretion has been shown to mediate leukocyte influx [14 ]. We and others have found that osteoblasts are a significant source of chemokines in M. tuberculosis and other bacterial bony infection [15 16 17 ].
In addition to chemokines, zinc containing matrix metalloproteinases (MMPs) is important in control of cellular migration [18
, 19
]. However, MMPs secreted in excess have the potential to cause enzymatic tissue destruction. MMP-9 secretion can be induced by proinflammatory cytokines including IL-1 and tumor necrosis factor (TNF-) [20
, 21
]. Elevated MMP gene expression and secretion have been described by ourselves and others in tuberculosis, and MMP-9 concentrations relate to local tissue damage [22
23
24
]. MMP-9, 92 kDa gelatinase, is quantitatively the most significant MMP-9 secreted by monocytic cells [25
]. There are no data about the role or cellular source of MMP-9 during bony M. tuberculosis infection.
In initial studies, M. tuberculosis induced CCL2 and CXCL8 mRNA expression in osteoblasts in a biphasic pattern with peak of gene expression at 2–4 h, which then decreased to reappear at 24 h [16
]. Such a pattern suggests the presence of paracrine bony networks. Networks are known to aid granuloma formation [26
], and in pulmonary tuberculosis, networks between respiratory epithelial cells and monocytes mediated by monocyte-derived IL-1 but not by TNF- may be more important in driving chemokine secretion than direct infection [27
]. Monocyte–monocyte networks in tuberculosis up-regulate MMP-9 secretion mediated by a TNF--dependent path [28
]. TNF- and IL-1 are thus potentially important regulators of networks in tuberculosis; osteoblasts can secrete and respond to these proinflammatory cytokines [29
, 30
].
We investigate the involvement of networks between osteoblasts and monocytes in control of chemokine and MMP-9 secretion in tuberculous osteomyelitis. Recruited monocytes are critical for granuloma formation in tuberculosis [31
]. Monocytes are particularly important in bone immune responses [32
, 33
]. We focused on the role of TNF- and IL-1, as they are critical for tuberculous granuloma development [34
, 35
], involved in other networks in tuberculosis, and important mediators of bone resorption [36
37
38
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-modified Eagle’s media (-MEM) growth media, supplemented with 10% fetal calf serum (FCS), 2 mM glutamine, 1% nonessential amino acids, and 10 µg/ml ampicillin in a humidified 5% CO2 atmosphere at 37°C. Cells were trypsinized and seeded in tissue-culture plates 24 h before use. Immediately before experiments, media were replaced with serum-free -MEM, supplemented with 2 mM glutamine, 1% nonessential amino acids, and 10 µg/ml ampicillin, which does not inhibit growth of M. tuberculosis at this concentration. Human peripheral blood monocytes were isolated from pooled leukocyte-enriched buffy coats (obtained from North Thames Blood Transfusion Service, Colindale, London, UK). Peripheral blood mononuclear cells were separated by density centrifugation on Ficoll-Paque (Amersham Pharmacia Biotech, Little Chalfont, UK), and monocytes were purified by adhesion to tissue-culture plastic for 2 h. Monocytes were maintained in RPMI-1640 medium (supplemented with 10% FCS, 2 mM glutamine, and 10 µg/ml ampicillin) overnight before performing experiments in serum-free media.
Mycobacterial culture
The virulent M. tuberculosis H37rv strain (obtained from the National Collection of Type Cultures, Colindale, UK) was maintained in Dubos medium supplemented with bovine serum albumin, sucrose, and sodium chloride (Becton Dickinson, San Diego, CA). Before use, mycobacteria were briefly sonicated and passed five times through a 22-gauge needle to generate a single-cell suspension, as confirmed by modified Kinyoun staining. Numbers of viable bacteria were quantified by colony counting on Middlebrook 7H10 plates.
Experimental model
To generate conditioned media from osteoblasts infected with M. tuberculosis (COBTB), MG-63 cells were cultured with M. tuberculosis at a multiplicity of infection (MOI) = 10, and tissue-culture supernatants were harvested 24 h later. COBTB was passed through a 0.2-µm filter to remove any viable bacteria. COBControl was obtained from unstimulated osteoblasts cultured for 24 h and was the negative control. Monocytes were stimulated with COBTB (diluted 1:5, 1:10, 1:50, or 1:100) or with COBControl for up to 48 h in serum-free RPMI for assessment of chemokine/MMP secretion and gene expression. The concentration of osteoblast-derived CXCL8 and CCL2 present within COBTB is represented by baseline data obtained at T = 0.
To investigate osteoblast-derived factors in cellular networking, cytokine neutralization experiments were performed. The role of TNF- and IL-1ß in autologous osteoblastic chemokine secretion was determined by first incubating the cells with neutralizing anti-TNF- antibody or IL-1 receptor antagonist (IL-1ra; both Peprotech, Rocky Hill, NJ) for 2 h before infection with M. tuberculosis (MOI=10). Next, the active factors within COBTB were investigated by preincubating with anti-TNF- antibody for 2 h before adding to the monocyte culture or culturing monocytes in the presence of IL-1ra for 2 h to block IL-1r before stimulation by COBTB. Previous work in this laboratory has confirmed that anti-TNF- antibody (50 µg/ml) neutralizes the effects of 10 ng/ml TNF- (in accordance with the manufacturer’s information) [27
]. Tissue-culture medium and RNA were harvested at 24 h for further analysis.
Enzyme-linked immunosorbent assay (ELISA)
Culture supernatants were harvested at time points indicated, passed through a 0.2-µm spin filter to remove viable organisms (direct infection experiments only), and frozen at –70°C for later chemokine assays. CXCL8 and CCL2 secretion were detected by ELISA using matched pairs of antibodies according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN). The lower limit of sensitivity of assays was 3 pg/ml.
RNA extraction and Northern blotting
Total RNA was extracted from cellular monolayers using Tri-ReagentTM (Sigma, Poole, UK), according to the manufacturer’s instructions. For Northern analysis, 20 µg RNA was electrophoresed on denaturing formaldehyde 1% agarose gels, transferred by capillary blotting to Hybond-N membrane (Amersham Pharmacia Biotech), and fixed by exposure to UV light using a UV Stratalinker 1800 (Stratagene, La Jolla, CA). Blots were prehybridized and then hybridized overnight with (
-32P) end-labeled oligonucleotide probes against CCL2 (probe cocktail purchased from R&D Systems), CXCL8 (30-mer), and the housekeeping gene ß-actin (42-mer) [39
]. Blots were autoradiographed at –70°C for 24–48 h and digitized before densitometric analysis. Between probings, blots were stripped by heating for 1 h at 65°C in probe remover [0.005 M Tris-HCl (pH 8), 0.002 M Na2HPO4, 0.1x Denhardt’s solution].
Gelatin zymography
Secreted MMP-9 was assessed by gelatin zymography using standard methodology [40
]. Briefly, cell-culture supernatants were mixed with 5x loading buffer [50 mM Tris-HCl (pH 7.6), 10% glycerol, 1% sodium dodecyl sulfate (SDS), and 0.01% bromophenol blue] before resolution on an 11% SDS gel, which had been impregnated with 0.12 mg/ml porcine gelatin (Sigma) and overlaid with a 4% stacking gel. Samples were run for 
3 h at 180 V and were washed for 1 h in 2.5% Triton X-100 before overnight incubation in collagenase buffer [50 mM Tris-HCl (pH 7.6), 0.2 M NaCl, and 5 mM CaCl2] at 37°C. The next day, gels were rinsed in water and stained with 0.2% Coomassie blue (Amersham Pharmacia Biotech) to reveal proteolytic activity as white bands on a dark background.
Densitometric and statistical analysis
All images were digitized and analyzed using NIH Image 1.58 (National Institutes of Health Research Services Branch, Bethesda, MD). Chemokine mRNA signal densitometry was normalized for total RNA using the housekeeping gene ß-actin. Results are expressed as change in chemokine mRNA compared with unstimulated cells (±SEM) and are from three independent experiments. Densitometric analysis of gelatinase activity was performed, and MMP-9 concentration was determined from standard curves, prepared from known quantities of recombinant human MMP-9 (32–8000 pg; Calbiochem, Nottingham, UK). Chemokine concentrations are expressed as pg/ml/4 x 105 cells, and data presented as mean ± SEM of experiments performed in triplicate on three separate occasions. Differences between groups were compared by unpaired t-test, as data were normally distributed. P< 0.05 is taken as statistically significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
secretion in M. tuberculosis-induced osteoblast chemokine secretion15-fold greater CCL2 secretion, similar to our previous findings [41
]. To investigate the role of proinflammatory mediators in autocrine networks, MG-63 cells were preincubated with neutralizing anti-TNF- antibody or IL-1ra. After 2 h, cells were infected with M. tuberculosis (MOI=10), and culture supernatants were harvested 24 h later. M. tuberculosis-induced CXCL8 secretion was reduced in a dose-dependent manner by anti-TNF- antibody, and concentrations decreased 50% by 50 µg/ml (from 6280±410 to 3070±220 pg/ml/4x105 cells; P<0.01; Fig. 1A
). CCL2 secretion was not affected by low anti-TNF- concentrations and only decreased 30% by maximal TNF- blockade (P<0.05; Fig. 1B
). IL-1 blockade did not affect CXCL8 or CCL2 secretion (Fig. 1C
and 1D)
.
 View larger version (36K): [in a new window] |
Figure 1. Inhibition of autologous TNF- secretion decreases M. tuberculosis-induced osteoblast CXCL8 and CCL2 secretion. MG-63 cells were incubated with increasing doses of neutralizing anti-TNF- antibody (A and B) or IL-1ra (C and D). After 2 h, the cells were stimulated with M. tuberculosis (MOI=10). Twenty-four hours later, culture supernatants were harvested and assayed by ELISA for CXCL8 (A and C) or CCL2 (B and D) protein. Results shown are representative of a triplicate experiment performed on two separate occasions, and chemokine levels are expressed as pg/ml/4 x 105 cells. *, P< 0.05; **, P< 0.01.
|
 View larger version (16K): [in a new window] |
Figure 2. COBTB induces concentration and time-dependent secretion of CXCL8 and CCL2 from human monocytes. (A) Freshly isolated monocytes were stimulated with COBTB (diluted 1:5–1:100) for 24 h before harvesting culture supernatants and assaying for CXCL8 and CCL2. (B) In kinetic experiments, culture supernatants from monocytes stimulated with COBTB (1:10) or COBControl were harvested at 0 h, 4 h, 24 h, and 48 h, and CXCL8 and CCL2 secretion was assessed by ELISA. Data are representative of a triplicate experiment performed on three separate occasions (mean±SEM).
|
and IL-1 were involved in paracrine networks. IL-1ra (0–200 ng/ml) significantly inhibited COBTB-induced CXCL8 secretion from COBTB-stimulated monocytes (P<0.01; Fig. 3A
). Anti-TNF- antibody reduced CXCL8 concentrations from 5480 ± 200 to 3350 ± 60 pg/ml/4 x 105 cells at 100 µg/ml in a less-marked but a dose-dependent manner. Neither IL-1ra nor anti-TNF- antibody altered CXCL8 mRNA accumulation, indicating a post-transcriptional mechanism of action (Fig. 3B
and data not shown).
 View larger version (21K): [in a new window] |
Figure 3. Effect of IL-1 blockade on COBTB-induced CXCL8 gene expression and secretion in human monocytes. (A) Monocytes were preincubated with IL-1ra (0–200 ng/ml) before being stimulated by COBTB (1:10). After 24 h, culture supernatants were harvested for CXCL8 ELISA. Data are representative of triplicate experiments performed on three separate occasions. (B) Cellular mRNA from control monocytes or COBTB-stimulated monocytes pretreated with 100 µg/ml anti-TNF- [antibody (Ab)] or 200 ng/ml IL-1ra and harvested at 24 h. Bar chart shows relative densitometry of CXCL8 mRNA expression corrected for ß-actin mRNA expression with representative Northern blots below. Data shown are representative of an experiment performed on three separate occasions (mean±SEM). *, P< 0.05; **, P< 0.01.
|
antibody (P<0.001; Fig. 4A
). IL-1ra had a relatively reduced effect compared with that on CXCL8 (Fig. 4B)
. These changes were also not associated with decreased CCL2 gene transcription (data not shown). However, not all monocyte responses to COBTB are regulated similarly. COBTB-induced secretion of MMP-9 was not down-regulated by blockade of either proinflammatory cytokine (Fig. 4C
and 4D)
. At concentrations used, neither anti-TNF- antibody nor IL-1ra affected cellular viability or basal secretion of monocyte-derived mediators CXCL8, CCL2, or MMP-9.
 View larger version (41K): [in a new window] |
Figure 4. COBTB-induced CCL2 but not MMP-9 secretion from monocytes is inhibited by TNF- and IL-1 blockade. Stimulating freshly isolated monocytes with COBTB (1:10) after 2 h treatment with anti-TNF- antibody (0–100 µg/ml) or IL-1ra (0–200 ng/ml) assessed the contribution of osteoblast-derived TNF- and IL-1 to secretion of CCL2 and MMP-9 from monocytes. At 24 h, culture supernatants were harvested, CCL2 secretion was assessed by ELISA (A and B), and MMP-9 secretion was analyzed by gelatin zymography (C and D). Data are mean ± SEM and are representative of triplicate experiments performed on three separate occasions. A representative zymogram is shown with MMP-9 data. *, P< 0.05; **, P< 0.01; ***, P< 0.001.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and IL-1 in networks in tuberculous osteomyelitis resulting in human monocyte CXCL8 and CCL2 secretion. TNF- but not IL-1 was involved in autocrine networks driving osteoblast CXCL8 and to a lesser extent, CCL2 secretion. Conditioned media from M. tuberculosis-infected MG-63 cells induces monocyte gene expression and secretion of CXCL8 and CCL2. TNF- and IL-1 are key mediators in this paracrine network, acting mainly post-transcriptionally. However, inhibiting these proinflammatory cytokines did not neutralize all the effects of COBTB. The effect was relatively specific, as monocyte MMP-9 secretion was not affected by blocking proinflammatory cytokine activity.
TNF- is secreted directly by activated osteoblasts, and autologous production accounts for about half of CXCL8 secretion during M. tuberculosis infection and less CCL2 secretion. As TNF- production is essential for development of mycobacterial granuloma [26
, 44
], it is probable that osteoblasts are involved in granuloma development in osteomyelitis. In contrast, blocking IL-1 did not affect autologous chemokine secretion in M. tuberculosis-infected osteoblasts, although IL-1 has other key roles in tuberculosis [35
]. Our data do not exclude IL-1 from having other actions within bone that influence cellular recruitment, as neutralization of TNF- and IL-1 decreases bone loss mediated by a reduction in cell recruitment [37
].
The fact that COBTB induced secretion of CXCL8 and CCL2 from monocytes suggests that cytokine networks involving monocytes may be key in control of chemokine secretion in tuberculous osteomyelitis. Monocyte-derived CXCL8 concentrations were comparable with those observed from directly infected osteoblasts, and the levels of secreted CCL2 were tenfold lower [16
]. Taken together with our previous data, current findings indicate that recruited monocytes stimulated by infected osteoblasts are a major source of CXCL8 during tuberculous osteomyelitis, whereas osteoblasts themselves are the major cellular source of CCL2. This is consistent with data detailing the importance of osteoblast-derived CCL2 in bone activity through correlation with monocyte recruitment temporally and spatially [14
]. Although the magnitude of CXCL8 and CCL2 secretion from COBTB-stimulated monocytes was similar, kinetics of secretion are distinct. High CCL2 concentrations were detectable at 4 h, which may reflect slightly higher constitutive CCL2 gene expression compared with CXCL8 gene expression in unstimulated, adhesion-purified monocytes. CXCL8 secretion followed gene expression and is likely to be tightly, transcriptionally regulated, as is characteristic in monocytes [45
]. Monocyte CXCL8 and CCL2 mRNA levels decreased 50%, 48 h after stimulation with COBTB (data not shown), whereas we showed that CCL2 gene expression was maintained for up to 72 h in direct infection [16
]. Such differences may reflect the continued action of osteoblasts regulating influx of monocytes, which differentiate during the immune response to M. tuberculosis.
Blocking activity of IL-1 in COBTB significantly reduced monocyte-derived CXCL8 concentrations. CCL2 secretion was less sensitive to the effects of proinflammatory cytokine blockade. The minimal effect on gene expression implies a post-transcriptional mechanism of action. In contrast, IL-1 blockade of pulmonary monocyte-epithelial cell chemokine networks activated by M. tuberculosis decreased CXCL8 secretion to baseline levels and was associated with significantly reduced gene expression [27
]. As proinflammatory cytokine inhibition only partially decreased COBTB-induced monocyte chemokine secretion, other factors secreted by osteoblasts during tuberculous infection may be involved in activation of monocytes and potential candidate molecules, including IL-6, IL-11, granulocyte macrophage-colony stimulating factor, and IL-18, and various chemokines [46
47
48
49
50
]. In murine tuberculosis, blocking local TNF- secretion results in compensatory up-regulation of other cytokines, leading to greater, more persistent inflammation [51
]. A similar mechanism might partly account for failure to down-regulate chemokine mRNA in COBTB-exposed monocytes treated with anti-TNF- or IL-1ra.
It is unexpected that TNF- and IL-1 in COBTB were not implicated in regulation of human monocyte MMP-9 secretion, although MMP-9 may be secreted by mouse osteoblasts in response to TNF- or IL-1ß [52
53
54
]. However, only low-level MMP-9 secretion was observed in contrast to direct infection of human monocytes by M. tuberculosis [23
, 55
]. Immunohistochemical analysis has demonstrated MMP-9 expression within human bone, localizing to sites of bone resorption, specifically to osteoclasts and mononuclear cells [56
, 57
]. Although we provide evidence for a paracrine network between osteoblasts and monocytes resulting in MMP-9 secretion, it is possible that networks involving osteoclasts or other cell types may be more important in secretion of MMPs involved in bone destruction in tuberculous osteomyelitis.
In summary, our data demonstrate that TNF is involved in autocrine networks leading to M. tuberculosis-induced osteoblast chemokine secretion. In addition, there is an IL-1- and TNF-dependent network by which osteoblasts activate monocytes, resulting in high-level CXCL8 as well as CCL2 secretion during infection with M. tuberculosis. However, this network does not alter monocyte MMP-9 secretion. Such networks may have a significant role in recruitment and activation of leukocytes in tuberculous osteomyelitis.
Received September 19, 2003; revised January 15, 2004; accepted January 17, 2004.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and IL-1ß selectively induce expression of 92-kDa gelatinase by human macrophages J. Immunol. 157,4159-4165[Abstract]
B-dependent network J. Immunol. 163,3936-3947-dependent monocyte networks J. Immunol. 171,5579-5586, LPS, and IL-1 ß Science 243,1467-1469, and IL-4R Bone 28,268-274[Medline]-inducing factor) is produced by osteoblasts and acts via granulocyte/macrophage colony-stimulating factor and not via interferon- to inhibit osteoclast formation J. Exp. Med. 185,1005-1012: lack of regulation by 17 ß-estradiol Endocrinology 130,2528-2534 in the pulmonary inflammatory response to Mycobacterium tuberculosis infection Infect. Immun. 70,2082-2089This article has been cited by other articles:
 |
|
 |
|
  M. V. Delpino, C. A. Fossati, and P. C. Baldi Proinflammatory Response of Human Osteoblastic Cell Lines and Osteoblast-Monocyte Interaction upon Infection with Brucella spp. Infect. Immun., March 1, 2009; 77(3): 984 - 995. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Tsiodras, G. Samonis, D. T. Boumpas, and D. P. Kontoyiannis Fungal Infections Complicating Tumor Necrosis Factor {alpha} Blockade Therapy Mayo Clin. Proc., February 1, 2008; 83(2): 181 - 194. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Taylor, J. M. Hattle, S. A. Dreitz, J. M. Troudt, L. S. Izzo, R. J. Basaraba, I. M. Orme, L. M. Matrisian, and A. A. Izzo Role for Matrix Metalloproteinase 9 in Granuloma Formation during Pulmonary Mycobacterium tuberculosis Infection Infect. Immun., November 1, 2006; 74(11): 6135 - 6144. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Campbell, D. Bennett, L. Nasir, E. A. Gault, and D. J. Argyle Disease- and cell-type-specific transcriptional targeting of vectors for osteoarthritis gene therapy: further development of a clinical canine model Rheumatology, June 1, 2005; 44(6): 735 - 743. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
