In Vivo Activity of Released Cell Wall Lipids of Mycobacterium bovis Bacillus Calmette-Guérin Is Due Principally to Trehalose Mycolates
http://www.100md.com
免疫学杂志 2005年第8期
Abstract
The hallmark of Mycobacterium-induced pathology is granulomatous inflammation at the site of infection. Mycobacterial lipids are potent immunomodulators that contribute to the granulomatous response and are released in appreciable quantities by intracellular bacilli. Previously we investigated the granulomagenic nature of the peripheral cell wall lipids of Mycobacterium bovis bacillus Calmette-Guérin (BCG) by coating the lipids onto 90-μm diameter microspheres that were mixed into Matrigel matrix with syngeneic bone marrow-derived macrophages and injected i.p. into mice. These studies demonstrated that BCG lipids elicit proinflammatory cytokines and recruit leukocytes. In the current study we determined the lipids responsible for this proinflammatory effect. BCG-derived cell wall lipids were fractionated and purified by liquid chromatography and preparative TLC. The isolated fractions including phosphatidylinositol dimannosides, cardiolipin, phosphatidylglycerol, phosphatidylethanolamine, trehalose monomycolate, trehalose dimycolate, and mycoside B. Trehalose dimycolate, when delivered to bone marrow-derived murine macrophages, induced the greatest secretion of IL-1, IL-6, and TNF- in vitro. Trehalose dimycolate similarly induced the greatest secretion of these proinflammatory cytokines in ex vivo matrices over the course of 12 days. Trehalose monomycolate and dimycolate also induced profound neutrophil recruitment in vivo. Experiments with TLR2 or TLR4 gene-deficient mice revealed no defects in responses to trehalose mycolates, although MyD88-deficient mice manifested significantly reduced cell recruitment and cytokine production. These results demonstrate that the trehalose mycolates, particularly trehalose dimycolate, are the most bioactive lipids in the BCG extract, inducing a proinflammatory cascade that influences granuloma formation.
Introduction
Despite a worldwide vaccination program and half a century of treatment with antibiotics, Mycobacterium tuberculosis, the causative agent of tuberculosis, remains one of the leading causes of death attributable to an infectious agent. The hallmark of infection with pathogenic Mycobacterium spp. is the formation of a granuloma. Classical tuberculoid granulomas consist of a nidus of infected phagocytes, surrounded by a layer of activated giant cells and epitheloid macrophages, with a mantle of lymphocytes and fibrosis at the periphery (1). The tuberculoid granuloma is a fascinating example of a balanced interplay between host and pathogen. The granuloma is essential for containment of tuberculosis, as activated inflammatory cells enclose the tubercle bacilli, preventing dissemination and transmission of the pathogen. However the structure of the granuloma sequesters viable bacilli in an environment distant from macrophage-activating T cells, thus appearing to protect the bacilli from elimination. The tuberculoid granuloma is also continually evolving, and the development of extensive necrosis and caseation in advanced disease allows for transmission of the bacteria. Hence, although the host may require the granuloma for containment, it is advantageous for the bacterium to modulate this structure to ensure both persistence, and eventually transmission.
Mycobacterial glycolipids have long been recognized to have immunomodulatory activity, notably the induction of granulomatous responses. Mycobacterial cell wall glycolipids induce foreign body-type granulomas in mice (2, 3), and we have shown that glycolipids of Mycobacterium bovis bacillus Calmette-Guérin (BCG)3 traffic within the host macrophage and spread to uninfected bystander macrophages (4). Released glycolipids included phosphatidylinositol dimannosides (PIM2), phosphatidylglycerol and cardiolipin (CL), phosphatidylethanolamine, trehalose monomycolate (TMM), trehalose dimycolate (TDM), and mycoside B (MycB) (5). These classes of lipids are shared with M. tuberculosis, and lipids from both species exert potent adjuvant-like effects. The mechanisms by which the mycobacterial lipids exert these effects have yet to be determined (5, 6, 7, 8).
A considerable body of data has been published detailing the in vivo activity of mycobacterial lipids. PIM2 and phosphatidylinositol hexamannoside are granulomagenic when delivered with alum, recruiting macrophages, neutrophils, and NKT cells. These properties are dependent upon the lipid portion of the molecule (9, 10). PIM2, when presented on 1-μm diameter microspheres, elicits TNF-, reactive nitrogen intermediates, and macrophage chemotactic protein-1 from IFN--primed bone marrow-derived macrophages (BMM) or thioglycolate-elicited peritoneal macrophages (5). TDM, also known as cord factor, can reproduce the pathophysiologic aspects of M. tuberculosis infection including granuloma formation and induction of proinflammatory cytokines. Doses of TDM as little as 1–5 μg injected i.v. are granulomagenic in the lungs of mice, and repeated doses are more pathogenic than a single, larger dose (11). TDM-associated pathology includes weight loss due to TNF--mediated cachexia, with associated hyperlipidemia, hypoglycemia and peritonitis with ischemic and hemorrhagic lesions in the gastrointestinal tract (12). TDM-coated biodegradable microspheres, when embolized into murine lungs, induce intense inflammatory infiltrates and IL-4, IL-6, TNF-, IL-12, IL-10, and IFN- (13). Hence, a variety of mycobacterial lipids have broad immunomodulatory capabilities that affect granuloma formation.
Microbial components or pathogen-associated molecular patterns are recognized predominantly by pattern-recognition receptors such as TLRs that play an important role in innate response to infection. Pattern-recognition receptors involved in recognizing M. tuberculosis include the mannose receptor, complement receptors, and TLRs (14). TLRs have an important role in triggering innate response to mycobacterial cell wall components, and the principal TLRs involved in recognition of mycobacterial lipids are TLR2 and TLR4. TLR2 binds mycobacterial lipoarabinomannan and phosphatidylinositol mannoside (PIM), and PIM also signals through TLR4 (14, 15, 16).
Although there are many studies detailing the bioactivity of individual mycobacterial cell wall components there has been no comparative analysis of the lipids known to be released within infected macrophages. The in vivo response of mice to the different lipid fractions was investigated using a lipid-coated bead/matrix model that was developed and described in an earlier study (17). In this model total BCG lipids induced proinflammatory cytokines and chemokines including TNF-, IL-1, IL-1, IL-6, macrophage chemotactic protein-1, IFN--inducible protein-10, and recruited polymorphonuclear and mononuclear leukocytes and lymphocytes. The current comparative study demonstrates that trehalose mycolates are the principal bioactive lipids of BCG when coated onto 90-μm diameter microspheres, eliciting cytokines in vitro and in vivo with associated recruitment of primarily neutrophils and macrophages. Intriguingly the mode of presentation of the lipids on large particles played a significant role in lipid bioactivity. We also show that the inflammatory response to TDM was not diminished in TLR2–/– or TLR4 defective mice, although cell recruitment and cytokine production in MyD88–/– mice was severely diminished indicating involvement of the TLR/IL-1R signaling pathway.
Materials and Methods
Mice
Age- and sex-matched wild-type (WT) mice including C57BL/6, C3H/HeN, and C57BL/6/129F2 strains were purchased from Charles River Breeding Laboratories and housed under specific pathogen-free conditions. Unless otherwise stated experiments were performed in C57BL/6 mice. Female 6- to 8-wk-old TLR4-deficient C3H/HeJ mice were purchased from The Jackson Laboratory. Male TLR2–/– mice were the kind gift of Dr. E. Unanue (Washington University, St. Louis, MO). MyD88-deficient mice on a C57BL/6/129F2 background were generated by Dr. S. Akira and provided by Dr. S. Ehrt (Cornell University, New York, NY).
Macrophage culture
Mice were anesthetized and sacrificed by cervical dislocation, and the femurs and tibias were harvested. Bone marrow was flushed from the bones with cold DMEM (Invitrogen Life Technologies), supplemented with L-929 cell-conditioned medium, heat-inactivated FCS (10%; Summit Biotechnology) and horse serum (5%; Sigma-Aldrich), L-glutamine (2 mM), sodium pyruvate (1 mM), 100 U/ml penicillin, 100 μg/ml streptomycin (Invitrogen Life Technologies). Bone marrow cells were cultured in bacterial grade 90-mm2 petri dishes (Valmark; Kord) at 37°C, 5% CO2 for 5 days, after which BMM were harvested in endotoxin-free PBS (Invitrogen Life Technologies) or expanded for an additional 3–5 days. BMM were used for in vitro or in vivo assays.
Mycobacterial culture and extraction of lipids
M. bovis BCG (Pasteur strain) was grown to late-log phase in Middlebrook 7H9 medium (Difco) supplemented with oleic acid, BSA-Fraction V, dextrose, NaCl and 0.05% Tween 80. Bacilli were washed extensively in 0.05% Tween 80/PBS followed by a detergent-free wash before extraction in chloroform and methanol as previously described (5). Briefly, the pellet was extracted twice in chloroform/methanol (2:1 v/v) for 15 min at 55°C with sonication. Bacterial debris was removed by passing the extract through a 0.2-μm Teflon filter, and hydrophilic contaminants were removed using a Folch wash. The extract was fractionated by liquid chromatography over silica gel 60 (EM Science) in increasing amounts of methanol in chloroform followed by preparative TLC on aluminum-backed silica gel 60 plates (EM Science). Individual lipids were scraped from the plate and extracted from the silica in chloroform: methanol (2:1 v/v) and purified further using reverse-phase chromatography through C18 Sep-pak columns (Waters). Lipids were detected on TLC plates using 50% ethanolic sulfuric acid and charring. Purified lipid fractions included those that had been identified previously: diphosphatidylglycerol CL, phosphatidylglycerol, phosphatidylethanolamine, PIM2a and PIM2b (with PIM2b having an additional acyl group attached to a mannose residue), TMM, TDM, and MycB (5). Lipids were resuspended in chloroform/methanol (2:1 v/v) at 10 mg/ml and stored at –20°C under nitrogen.
In vitro macrophage assays
BMM were plated overnight in 24-well plates at 2 x 105 per well in L929-supplemented DMEM. Sterile polystyrene microspheres (90-μm diameter; Polysciences) were coated passively in an excess of total BCG lipid extract, purified lipid fractions, or bovine-derived phosphatidylglycerol (Sigma-Aldrich) in PBS using alternate cycles of vortexing and sonication at 55°C. Coated microspheres were washed and resuspended in endotoxin-free PBS at 2% solids. For lipid comparison in vitro, coated microspheres were added to macrophage monolayers at an equivalent ratio to that used in the in vivo model, and noncoated beads or nonpulsed macrophages were included as controls. For bead size assays, an equivalent surface area of microspheres (1, 2, 3, 4.5, 6, 10, 25, and 90 μm diameter) was coated with an excess of TDM as earlier described. Microspheres were added to macrophages prepared as previously discussed, over a dose value of 0.5x, 1x, and 2x with 1x equivalent to a total bead surface area of 5 x 107 μm2. Cells were incubated for 48 h at 37°C in 5% CO2, after which supernatants were collected and stored at –70°C until analysis.
Delivery and harvest of lipid-bearing matrices
Polystyrene microspheres (90 μm in diameter) were coated with lipid and mixed (6.7% solids) with syngeneic BMM (1 x 107/ml) in ice-cold, phenol red-free, growth factor-reduced Matrigel (BD Biosciences), a semidefined gel composed of murine laminin, collagen (type IV), entactin, and minimal concentrations of growth factors (epidermal growth factor, nerve growth factor, platelet-derived growth factor, and TGF-). Liquid matrix (300 μl) was injected i.p. into mice upon which it polymerized at body temperature. At various times postinjection mice were sacrificed and matrices were recovered and washed in endotoxin-free PBS (Invitrogen Life Technologies).
Histology
Matrices were recovered at various times postinjection and fixed in 4% neutral-buffered paraformaldehyde. Matrices were mounted in Histogel (Richard-Allan Scientific) before routine histologic processing and paraffin embedding. Sections (4-μm thickness) were stained with H&E. Representative micrographs were captured by an AxioCam on a Zeiss Axioskop II.
Cytokine ELISA
Peritoneal matrices were harvested at various times postinjection, and cultured ex vivo for an additional 48 h. Briefly, preweighed matrices (30 mg) were cultured in DMEM containing FCS, L-glutamine, sodium pyruvate, penicillin/streptomycin at 37°C 5% CO2. Culture supernatants were collected and stored at –70°C until analysis. TNF-, IL-1, IL-2, IL-4, IL-6, and IFN- were measured using BD OptEIA ELISA kits (BD Pharmingen). Vascular endothelial growth factor (VEGF) was measured using a Quantikine M ELISA kit (R&D Systems). Supernatants from individual mice (n = 3) were assayed and normalized to the weight of the matrix.
Flow cytometry
Peritoneal matrices were recovered, weighed, incubated briefly in cold BD Cell Recovery solution (BD Biosciences) and dispersed by passage through a 70-μm mesh. The numbers of viable cells were counted in trypan blue to calculate the number of viable cells per milligram of matrix. For immunophenotyping, cells were blocked in FACS buffer (ice-cold PBS with 10% heat-inactivated FCS and 5% normal mouse serum; Jackson ImmunoResearch Laboratories) and anti-FcRIII/RII (clone 2.4G2; BD Pharmingen) for 15 min. Cells (from individual mice (n = 3) or pooled samples) were subsequently stained in FACS buffer using fluorophore-conjugated Ab for 45 min at 4°C, and propidium iodide at 3 μg/ml (BD Pharmingen) was added during the final 10 min. Stained cells were washed in FACS buffer and resuspended in 10% FCS in PBS for immediate flow cytometric analyses.
Live leukocytes were gated upon as propidium iodide-negative events with a scatter profile typical of a mixed leukocyte population. Within this gate, neutrophils were identified as small Ly-6GhighF4/80– cells, macrophages were large Ly-6GvariableF4/80+I-A/I-E+ cells, and eosinophils were small, granular F4/80midI-A/I-E– events. B cells were agranular CD19+ events, T cells were small, CD3+ events, and NK cells were small, agranular CD49b+ (CD3+ or CD3–) events. Abs purchased from Caltag Laboratories included: anti-Ly-6G (clone RB6–8C5), anti-macrophage marker (clone F4/80), anti-CD49b (clone DX5), and anti-CD19 (clone 6D5). Abs purchased from BD Pharmingen included: anti-I-A/I-E (clone 2G9), anti-CD3 (clone 145-2C11), and isotype control Abs. Samples were analyzed on a FACSCalibur equipped with two lasers (488 and 633 nm) and analyzed using CellQuest software (BD Biosciences). Data were presented as the number of viable leukocytes for each subset per milligram of matrix (percentage of viable leukocytes x number of viable cells per milligram of gel).
Statistical methods
Cytokine measurements and calculations of cell numbers were performed on samples from individual mice and the SD of the mean values (n = 3) are presented. The statistical significance of differences in the means of samples relative to phosphatidylglycerol-induced responses was calculated using a Student’s paired t test (p < 0.05 and p < 0.005). Statistical differences in cell recruitment and cytokine production from TDM-bearing matrices of MyD88–/– and WT animals are indicated (p < 0.005).
Results
BCG lipid fractions elicit proinflammatory cytokines when delivered to BMM in vitro
Previous studies determined that total BCG lipids, when delivered i.p. into mice on 90-μm diameter beads with syngeneic BMM in the matrix, induced both leukocyte recruitment and proinflammatory cytokine production (17). To determine which of the lipids were proinflammatory, total BCG lipids were fractionated and the principal species were purified (Fig. 1). These lipids were coated onto 90-μm diameter beads and screened on BMM in vitro. Purified lipids included PIM2a (diacyl), PIM2b (triacyl), diphosphatidylglycerol CL, phosphatidylglycerol, phosphatidylethanolamine, TMM, TDM, and MycB. These had been identified previously as the major lipid species released from intracellular BCG and had been shown to traffic to neighboring, uninfected bystander macrophages (5).
FIGURE 1. Fractionated lipids of M. bovis BCG. Total peripheral cell wall lipids were extracted from cultures of BCG in 2:1 (v/v) chloroform/methanol. Lipids were fractionated by liquid chromatography and preparative TLC. A, Resolved in chloroform/methanol (90:10 v/v) to distinguish TMM, TDM, and MycB. B, Resolved in chloroform/methanol/water (65:25:4 v/v) to separate diacyl PIM2a, triacyl PIM2b, CL/diphosphatidylglycerol, and phosphatidylethanolamine (PE). Lipids are compared (A and B) with total BCG peripheral cell wall lipid extract (Total).
Secretion of IL-1, IL-6, and TNF- from BMM was elicited by some individual lipid fractions or by the total BCG lipid extract (Fig. 2). TDM was the principal inducer of IL-1, eliciting higher amounts than total BCG lipids (Fig. 2A). Several lipids induced IL-6 significantly above controls, including CL, TMM, TDM, PIM, and total BCG lipids (Fig. 2B). The trehalose mycolates and total BCG lipids were the strongest inducers of TNF-, although substantial amounts of the cytokine were detected from CL- and PIM-treated cells (Fig. 2C). PE or MycB failed to induce these cytokines. Consequently, CL, PIM2b, TMM, and TDM were selected for further analyses in vivo. PIM2b was selected for further investigation because it was slightly more stimulatory than PIM2a.
FIGURE 2. BCG-derived trehalose mycolates induce cytokines from BMM when delivered on 90-μm polystyrene microspheres. Lipid-coated polystyrene microspheres (90-μm diameter) were overlaid onto confluent monolayers of C57BL/6 BMM. Beads were added to macrophages at a ratio equivalent to that used for in vivo matrices. phosphatidylglycerol-coated beads (PG), noncoated beads (Bead), and nontreated macrophages () were included as controls. Cells were incubated for 48 h, and supernatants were analyzed by sandwich ELISA for IL-1 (A), IL-6 (B), and TNF- (C). Data from triplicate samples from a representative experiment are shown as mean ± SD, and the statistical significance is shown relative to phosphatidylglycerol (*, p < 0.05; **, p < 0.005).
Previous studies had determined that PIM2a or PIM2b, when presented on 1-μm diameter beads to IFN--primed BMM or thioglycolate-elicited peritoneal macrophages, was the principal proinflammatory lipid of the total BCG lipid extract (5). This finding was in contrast with the current study, which indicated that trehalose mycolates were the principal inflammatory lipids when presented on 90-μm diameter beads. To examine the effect of bead size on lipid bioactivity, the same batch of purified peripheral cell wall lipids of BCG were presented to unprimed or IFN--primed BMM on either 1-μm diameter or 90-μm diameter beads, and the secretion of TNF- was assayed. These studies recapitulated earlier results with PIM2 being the principal inducer of TNF- when presented on 1-μm diameter. beads (5). In contrast, trehalose mycolates elicited no TNF- on the smaller particles but were the most proinflammatory fraction when presented on 90-μm diameter beads to unprimed BMM (data not shown). To determine the bead diameter necessary for trehalose mycolate bioactivity, TDM was presented to unprimed BMM on microspheres of increasing size from 1- to 90-μm diameter, at a range of concentrations (Fig. 3). TNF- was elicited by beads of 10-μm diameter, which were phagocytosed; however cytokine secretion was maximal and equivalent for larger (25- or 90-μm diameter) beads to which the macrophages attached but did not internalize. These results indicate that the bioactivity of trehalose mycolates is highly dependent on the mechanism of presentation.
FIGURE 3. The bioactivity of TDM is dependent upon the size of the particle. Trehalose mycolate-coated polystyrene microspheres of increasing diameter (1, 2, 3, 4.5, 6, 10, 25, and 90 μm) were overlaid onto confluent monolayers of BMM. The numbers of beads delivered to the monolayers were normalized to bead surface area, and beads were added over a range with 1x being equivalent to a total bead surface area of 5 x 107 μm2/well. Cells were incubated for 48 h, and the supernatants were analyzed by ELISA for TNF-. Data from triplicate samples from a representative experiment are shown as mean ± SD.
Trehalose mycolates elicit proinflammatory cytokines in peritoneal matrices ex vivo
CL, TMM, TDM, and PIM2b were delivered to mice in matrices, and the inflammatory response in the recovered matrices was followed over 12 days. C57BL/6 mice were injected with matrices containing syngeneic BMM and 90-μm diameter beads coated with fractionated lipids, total lipids, or bovine phosphatidylglycerol. Matrices were harvested from mice after 14 h and after 4, 7, and 12 days, weighed, and cultured for an additional 48 h to compare levels of cytokine secretion. Minimal cytokine secretion had been measured from ex vivo cultured matrices that contained phosphatidylethanolamine, PIM2a, or MycB at 14 h and at 4 days, indicating that these lipids were not proinflammatory (data not shown). In contrast, TDM induced high levels of IL-1, IL-6, and TNF- (Fig. 4) with notably enhanced secretion of IL-1 that peaked at 14 h (Fig. 4A). TMM also induced IL-1, albeit at lower levels at 14 h, which declined after 4 days. Total BCG lipids induced moderate levels of IL-1 with similar kinetics to TDM. In contrast to the trehalose mycolate containing fractions, PIM2b and CL failed to elicit IL-1 at levels greater than the control lipid, phosphatidylglycerol. There was sustained secretion of IL-6 in response to TMM, TDM, and total BCG lipids (Fig. 4B). Levels of IL-6 induced by CL and PIM2b were similar to phosphatidylglycerol, increasing slowly over 12 days. TMM and TDM induced comparably high secretion of TNF- that peaked within 14 h (Fig. 4C). Previously, we had noted that BCG lipid-bearing matrices tended to attach to peritoneal tissue and vascularize (17). We investigated whether VEGF, a powerful proangiogenic factor was involved in the response to the bacterial lipids. Secretion of VEGF from TMM, TDM, and total BCG lipid matrices peaked at 14 h, but by 4 days secretion of VEGF was comparable in all matrices (Fig. 4D). There were no differences in IL-2, IL-4, or IFN- production stimulated by any of the lipid fractions, although low levels were detected at 7 days (data not shown). These data establish that trehalose mycolates are the major proinflammatory lipids in the total BCG extract.
FIGURE 4. BCG-derived trehalose mycolates also induce cytokines in peritoneal matrices. Mice were injected with matrices containing syngeneic C57BL/6 BMM, and the purified BCG lipids CL, PIM2b, TMM or TDM, total BCG lipids (Total), or bovine-derived phosphatidylglycerol (PG), and the matrices were recovered at the indicated times and cultured for an additional 48 h. Supernatants were collected and analyzed by ELISA. Levels of IL-1 (A), IL-6 (B), TNF- (C), and VEGF (D) were normalized to the weight of matrix. Data from triplicate samples from a representative experiment are shown as mean ± SD, and significantly different levels from phosphatidylglycerol-bearing matrices are indicated (*, p < 0.05; **, p < 0.005).
Trehalose mycolate-bearing matrices induce marked recruitment of cells, most notably of neutrophils
The recruitment of leukocytes to matrices bearing CL, PIM2b, TMM, or TDM was compared with those containing total BCG lipid or phosphatidylglycerol. BCG lipids recruited significantly more cells than phosphatidylglycerol, as shown previously (17); however, the trehalose mycolates recruited even greater numbers of viable cells through 12 days (Fig. 5A). TMM recruited more cells than total BCG lipids at 14 h and at 4 days before falling to similar levels as controls at 7 days. CL and PIM2b recruited similar numbers of cells to phosphatidylglycerol throughout the time course. To determine which leukocyte subsets were recruited to the matrices, cell suspensions were stained for flow cytometry. The principal leukocytes recruited to trehalose mycolates were neutrophils (Ly-6GhighF4/80–)(Fig. 5B). Neutrophil recruitment kinetics mirrored the kinetics of total cell numbers for the various lipid fractions. Macrophage (Ly-6GvariableF4/80+I-A/I-E+) numbers were steady throughout the 12-day time course with little difference between the fractions (Fig. 5C). TDM recruited significantly higher numbers of eosinophils (forward scatterlow, F4/80mid, I-A/I-E–) at 14 h (Fig. 5D). There were no significant differences in the numbers of NK cells (side scatterlow, CD49b+), T cells (CD3+), or B cells (CD19+) recruited to the various lipid fractions, and numbers were low relative to granulocytes (data not shown). These data indicate that the enhanced cell recruitment observed in trehalose mycolate-containing matrices is due predominantly to neutrophils.
FIGURE 5. Trehalose mycolates recruit enhanced numbers of neutrophils to peritoneal matrices. Mice were injected (i.p.) with matrices that contained syngeneic C57BL/6 BMM with BCG-derived CL, PIM2b, TMM, TDM, Total lipids, or bovine phosphatidylglycerol (PG). Single cell suspensions from preweighed matrices were harvested and counted in trypan blue to calculate the total number of viable cells per milligram of matrix (A). Cells were also stained for flow cytometry, excluding dead cells from the analysis using propidium iodide. The percentages of viable (propidium-iodide negative) leukocyte subsets were converted to total cell numbers. The number of neutrophils (Ly-6GhighF4/80–) (B), macrophages (Ly-6GvariableF4/80+I-A/I-E+) (C), and eosinophils (side scatterhighF4/80midI-A/I-E–) (D) are indicated. Data from triplicate samples from a representative experiment are shown as mean ± SD, and significantly different levels from phosphatidylglycerol-bearing matrices are indicated (*, p < 0.05; **, p < 0.005).
Gross and histologic features of BCG lipid fraction granulomas
To analyze tissue development in response to the various lipid fractions, matrices containing CL, PIM2b, TMM, TDM, total BCG lipid, or phosphatidylglycerol were recovered for histologic examination. In gross terms, PIM2b and CL matrices resembled phosphatidylglycerol matrices, consisting of small masses floating free within the peritoneal cavity (data not shown). Generally phosphatidylglycerol, PIM2b, and CL matrices were fractured into several individual pieces. Opacity and firmness of the masses increased with time reflecting the accumulation of cells and debris. Due to substantial cell recruitment, TMM and TDM matrices exhibited marked opacity and firmness at 14 h, equivalent to that seen in a 7-day total BCG matrix. Additionally, all of TDM and TMM matrices adhered to peritoneal tissue by 4 days postinjection. Matrices typically adhered to adipose tissue within the peritoneal cavity, usually to the broad ligament, omentum, and mesentery. Once adhered, the organization of the matrices rapidly changed with recruited cells migrating in from the adherence margin where accumulations of leukocytes were observed. Interestingly adherent matrices vascularized, exhibiting substantial vessel growth and hemorrhage. Trehalose mycolate matrices were more cohesive than all other matrices and usually formed one large mass within the peritoneum. Trehalose mycolate matrices, particularly TMM also induced hemorrhagic ascites, most markedly at 4 days postinjection (data not shown). Trehalose mycolate matrices were well vascularized and by 7 days large areas of hemorrhage were visible, particularly in the TMM matrices. In some animals receiving trehalose mycolate matrices, hyperemia and vasculature hypertrophy developed in the vessels of the peritoneal body wall. Splenomegaly and peritoneal lymphadenopathy were also observed and were maximal at 7 days (data not shown).
Fig. 6 displays photomicrographs of the lipid-containing matrices over a 12-day time course. CL, phosphatidylglycerol, and PIM2b (PIM2b shown in Fig. 6, A–D) matrices remained relatively acellular throughout the 12-day time course. These lipid fractions recruited a mixed leukocyte population, and cellular infiltrates were restricted to the margins of the matrix and to surrounding the coated beads. In comparison total BCG lipid matrices exhibited early neutrophil recruitment at 14 h (Fig. 6E) and peak lymphocyte recruitment at 7 days (Fig. 6G). Cell numbers within the total BCG lipid matrices were intermediate between the phosphatidylglycerol-like fractions and the trehalose mycolates. Histology confirmed that cell recruitment to matrices containing trehalose mycolates was more profound. By 14 h substantial neutrophil recruitment had occurred in matrices containing TMM (Fig. 6I) or TDM (Fig. 6M). In contrast to total BCG lipids, neutrophil recruitment to both TMM (Fig. 6J) and TDM (Fig. 6N) was sustained. By 7 days, spaces were evident in the TMM and TDM matrices, likely due to degradation of the matrix by the recruited cells; this was particularly evident in the vicinity of the beads (Fig. 6, K and O). In response to all lipid fractions, multinucleated giant cells were present peripheral to the beads by 12 days (indicated by arrows, Fig. 6, D and P).
FIGURE 6. Histology of lipid-bearing matrices. Matrices containing lipid-coated beads and syngeneic C57BL/6 BMM were injected into mice and harvested at 14 h and at 4, 7, or 12 days. Paraformaldehyde-fixed paraffin-embedded sections were stained with H&E. Photomicrographs (x200) were taken of the most cellular regions within each section. Circular holes present within the amorphous matrix occur where coated beads were lost during processing. Sections containing PIM2b (A–D) exhibited similar cellularity to phosphatidylglycerol (PG) and CL (data not shown). Sections with TMM (I–L) or TDM (M–P) were more cellular than those with total BCG lipid (E–H), exhibiting substantial neutrophil recruitment. Note that TMM promoted greater hemorrhage within the matrix than TDM, and this was noted both grossly and histologically.
Cellular response to trehalose mycolates is independent of TLR2 or TLR4 but dependent on MyD88
Responses to most pathogen-associated molecular patterns characterized to date signal through the TLRs, and TLR2 and TLR4 bind mycobacterial constituents. To determine whether the cellular response to TDM was dependent on TLR2 or TLR4, cytokine production and cell recruitment to TDM-bearing matrices was assessed in mice deficient in these receptors. TDM-bearing matrices with syngeneic BMM were injected i.p. into TLR4- or TLR2-deficient mice or WT controls, and the responses were compared against phosphatidylglycerol-bearing matrices. The matrices were assessed for cell recruitment and cytokine production at 14 h postinjection. Minimal cell recruitment was seen in the phosphatidylglycerol matrices in any of the mice; in contrast, there were substantial and comparable levels of cell recruitment to TDM (Fig. 7, A and E) in TLR-deficient and WT mice. Similarly there was no defect in the production of IL-1, IL-6, or TNF- in either TLR4-deficient (Fig. 7, B–D) or TLR2-deficient (Fig. 7, F–H) mice, and there were no histologic differences compared with WT mice (data not shown).
FIGURE 7. Responses to TDM by TLR2-, TLR4-, and MyD88-deficient mice. Matrices containing either TDM or phosphatidylglycerol (PG), and syngeneic BMM were injected (i.p.) into TLR4-, TLR2- or MyD88-deficient mice or WT controls (MyD88–/– phosphatidylglycerol group was not determined (N.D.) due to scarcity of gene-deletion animals). Matrices were recovered 14 h postinjection and analyzed for cell recruitment and cytokine production as previously described. TLR4-deficient mice exhibited no defects in response to TDM in terms of cell recruitment (A), or cytokine production (B–D) relative to phosphatidylglycerol controls. Likewise, TLR2-deficient mice exhibited no defect in cell recruitment (E) or cytokine production (F–H) to TDM, relative to controls. In contrast MyD88-deficient mice had profound defects in cell recruitment (I), and cytokine production (J–L) compared with WT. Data from a representative experiment are presented as mean of matrices from individual mice ± SD. Statistical significance is shown relative to phosphatidylglycerol in WT for each group (*, p < 0.05; **, p < 0.005). Statistical differences in cell recruitment and cytokine production from TDM-bearing matrices of MyD88–/– and WT animals is indicated (, p < 0.005). There were no statistically significant differences between TLR-deficient and WT animals.
As a preliminary determination of whether other TLRs could be implicated in the recognition of TDM, we examined mice deficient in the common TLR-signaling adapter MyD88. TDM-bearing matrices from MyD88–/– mice were assayed for cell recruitment (Fig. 7I) and secretion of IL-1, IL-6, or TNF- (Fig. 7, J–L). MyD88–/– mice exhibited a profound defect in cell recruitment and cytokine production in TDM-bearing matrices compared with WT animals. These results indicate that the response to TDM, although independent of TLR2 or TLR4, is attenuated by the absence of the TLR/IL-1 adaptor MyD88.
Discussion
This study compared the relative biological activities of individual cell wall lipids that are released by M. bovis BCG into host macrophages. Comparison of seven principal peripheral cell wall lipids showed that trehalose mycolates were the most potent proinflammatory agents. When coated onto 90-μm beads and incubated with macrophages in vitro or inoculated into mice within matrices, trehalose mycolates induced substantial levels of IL-1, IL-6, TNF-, and VEGF with marked recruitment of neutrophils in vivo.
The inflammatory and toxic natures of trehalose mycolates have been appreciated for half a century. The physiologic dose of TDM from pulmonary mycobacterial infection in mice has been estimated to be 5 μg, and mice respond to doses of 1–5 μg (11). In our study, the beads were coated in excess of TDM, delivering an estimated 3–4 μg to each mouse. Mycobacterium smegmatis-derived TDM, delivered as an oil-in-water emulsion, induces IL-1, IL-6, and TNF- expression in a lung granuloma model with no detectable changes in mRNA for IFN-, IL-2, IL-4, IL-10, or IL-12 (18). Similarly, in our model, trehalose mycolates did not induce IL-12p70, the TH1 cytokine, IFN-, or the TH2 cytokines IL-4 or IL-10. Cytokines induced by trehalose mycolates contribute to several aspects of tuberculosis. Levels of TNF- induced by infection or TDM correlate with inflammation at the tissue level (18). TNF- induced by TDM after repeated injections also induces cachexia, a prominent clinical feature in end-stage tuberculosis (12). The weight of mice was not monitored in our study although none exhibited any sign of physical distress. A striking feature of trehalose mycolates matrices was sustained induction of high levels of IL-6. BCG-infected BMM also release IL-6, which inhibits proliferation of 3A9 T cells on hen-egg lysozyme-pulsed APCs (19). M. tuberculosis-infected macrophages secreted IL-6,which inhibited the response of uninfected macrophages to IFN- (20). Similarly, the high levels IL-6 induced by trehalose mycolates may suppress T cell and macrophage responses in the microenvironment of the granuloma. Total BCG lipids, TMM or TDM induced similar levels of IL-6 suggestive of synergistic induction by total BCG lipids because the total concentration of TMM and TDM combined in total BCG lipids was less than in matrices containing the purified fractions.
The striking features of trehalose mycolate-bearing matrices were prominent neutrophil recruitment accompanied by adherence and neovascularization. Neutrophil recruitment has been reported in other models of TDM-elicited inflammation (21, 22). Silva (3) has described cell recruitment to BCG or charcoal particles that were coated with TDM or TMM and injected (i.v.), subsequently lodging in the lungs. In contrast to our study, a moderate mononuclear infiltrate was noted, whereas intense neutrophil infiltrates were the response to glycerol monomycolate and glucose monomycolate. Macrophage recruitment to the matrices remained constant irrespective of the nature of the lipid.
The role of neutrophils in immunity to tuberculosis remains controversial. Neutrophil depletion during early mycobacterial infection exacerbates disease and dysregulates granuloma formation (23, 24, 25, 26). However neutrophils are associated with tissue damage and necrosis in late disease and are a feature of granulomas in gene-deletion models associated with increased susceptibility to tuberculosis (27, 28, 29, 30, 31, 32, 33). Neutrophils release mediators that likely contributed to the angiogenesis in the trehalose mycolate-bearing matrices, evident at 4 days postinjection (34). Several leukocyte products play a role in angiogenesis, including VEGF, IL-1, IL-6, IL-8, basic fibroblast growth factor, reactive nitrogen intermediates, TNF-, and TGF- (35, 36, 37, 38, 39). An appreciation for the role of neutrophils in angiogenesis is growing; neutrophils are sources of proangiogenic mediators including VEGF, IL-8, hepatocyte growth factor, and platelet-activating factor, and they can recruit additional leukocytes, like macrophages that contribute further to the proangiogenic cascade (34, 40). VEGF, an acute-phase proinflammatory response cytokine, increases vascular permeability and monocyte recruitment. Sakaguchi et al. (21) demonstrated TDM-enhanced neovascularization of murine air pouches with Rhodococcus-derived TDM, and neovascularization was attributed to the presence of VEGF-expressing neutrophils and macrophages. TDM from M. tuberculosis also induces neovascularization in a rat cornea model by cytokine-dependent mechanisms (35). Neutralizing Abs to VEGF or IL-8 substantially inhibited neovascularization, whereas anti-TNF- produced a partial inhibition; however, the cellular source of the angiogenic cytokines was not determined. In our model, trehalose mycolates or total BCG lipids induced an early burst of VEGF secretion from ex vivo matrices, coincident with the enhanced arrival of neutrophils (Fig. 4D), suggesting a role for both VEGF and neutrophils in the vascularization of adherent matrices. Although the granule protein eosinophil peroxidase has inhibitory activity toward M. tuberculosis H37Rv in vitro, the presence of eosinophils in pulmonary models of tuberculosis is considered detrimental as they are reflective of a harmful TH2 cytokine profile and contribute to tissue necrosis (41, 42, 43). Eosinophil recruitment was noted in response to trehalose mycolates, and likely reflects a lack of TH1-bias in the cytokine profile of matrices.
It has been reported that repeated injections (i.p.) of TDM-coated beads induces abdominal pathology very similar to that observed in TDM- or TMM-bearing matrices, including prominent peritoneal adhesions, hyperemic abdominal vessels, abdominal lymphadenopathy, splenomegaly, and a bloody ascites. Adhesions in the peritoneal cavity upon injection of TDM-coated beads initially consisted of neutrophils, and later of macrophages and fibroblasts (44). In our model, most lipid-bearing matrices floated freely in the peritoneal cavity, with the recruited cells coming from the surrounding peritoneal exudate. However, all of the TMM and TDM matrices were adherent by 4 days postinjection, a frequency higher than noted previously for matrices containing total BCG lipids (17). Comparison of adherent and free-floating total BCG lipid-bearing matrices from the same animal revealed that adherent matrices were more cellular, with infiltrates penetrating further into the matrix along the zone of adherence as well as in the regions of neovascularization. Adherence is significant in this model as it allowed recruitment of cells directly from the inflamed tissues, reflecting the mode of cell recruitment to tuberculoid pulmonary granulomas. Adherence of all trehalose mycolate matrices in the present study suggests that these lipids are responsible for this phenomenon.
The mechanism that promotes adherence of trehalose mycolate-bearing matrices is unknown. Retzinger et al. (22) determined that fibrinogen is a potent cofactor in TDM monolayers on coated beads, inducing more rapid and increased footpad swelling. There was thrombin-like procoagulant activity in the peritoneal cavities of mice injected with TDM-coated beads, and maximum biological activity occurred when TDM-coated beads were coated with blood plasma, which contains fibrinogen. It was postulated that thrombin-dependent conversion of fibrinogen to fibrin in situ was necessary for the biological activity of TDM monolayers. Any exposed hydrophobic residues are likely to be coated with fibrinogen in vivo, and conversion to fibrin and formation of adhesions may explain the adherence of matrices to tissues in our model. It is possible that this procoagulant activity enhances the mechanical adhesiveness of trehalose mycolate-bearing matrices, and we are currently investigating this in the matrix model.
Mechanisms of cellular activation by trehalose mycolates have yet to be determined. M. tuberculosis or its cell wall components can stimulate cells via TLRs (45, 46, 47). TLR2 and TLR4 are the principal TLRs reported to mediate recognition of M. tuberculosis, however neither TLR2- nor TLR4-deficient mice show a defect in cytokine production or leukocyte recruitment in response to TDM. The literature concerning the roles of TLR in tuberculosis infections is varied and contradictory (31, 33). TLR2 appears to play a minor role although TLR2-deficient mice succumb more readily to high dose challenge with M. tuberculosis (48). TLR4-deficient mice show a chronic neutrophilic pneumonia with reduced production of proinflammatory cytokines following challenge with M. bovis BCG (15, 49). In contrast, experiments with MyD88, the adaptor molecule required for signaling by TLRs and IL-1R, have in most instances revealed an increased susceptibility to mycobacterial infection (33). In the current study, MyD88–/– mice were defective in cell recruitment and cytokine production to trehalose mycolates compared with WT controls. This suggests that trehalose mycolates activate macrophages using a different TLR-mediated or IL-1 family receptor-mediated pathway than TLR2 or TLR4. Further analysis is required to determine the receptor(s) involved in the cellular recognition and response to TDM.
It is interesting to note that in a previous in vitro comparative study, we showed that PIMs were the most stimulatory lipids when presented on 1-μm beads to BMM, whereas the same numbers of trehalose mycolate-coated beads did not induce cytokines (5). This disparity was so striking that we compared responses to the same batch of lipid coated onto either 1- or 90-μm diameter beads (data not shown). Trehalose mycolates presented on 90-μm beads, but not 1-μm beads, induced inflammatory cytokines in vitro and in vivo. When a range of bead sizes was examined TDM-coated 10-μm diameter beads that could be phagocytosed elicited some TNF-, but secretion was maximal and equivalent for 25- and 90-μm diameter beads to which the macrophages attached. These data suggest that factors such as the number or receptors ligated, the duration of receptor/ligand interaction, or the formation of an adhesive plaque might modulate the cell response to TDM. These results recapitulate the elegant studies of Hunter and colleagues (22, 50) who documented that the biological activity of TDM was dependent on its presentation as a planar surface, in a monolayer, on a large particle, or in an oil/water interface. Given these structural requirements for biological activity it is pertinent to ask what relevance this might have in the course of disease. During the late stages of tuberculosis in guinea pigs, rabbits, and humans there is loss of granuloma containment, and breakdown of cavitary lesions, with uncontrolled extracellular mycobacterial replication, and the accumulation of cellular and lipid debris that is associated with neutrophil recruitment (51). Mice do not typically form the classical central caseous necrosis of pulmonary granulomas, however accumulation of lipid and cellular debris, neutrophil recruitment and fibrosis are common features of end-stage tuberculosis (52, 53, 54). The pathologic effects of trehalose mycolates may be predominant in the decompensating phase of tuberculosis, when uncontrolled extracellular bacillary growth and bacterial pellicle formation, with the accumulation of lipid debris present the lipid in a more planar configuration, similar to that in our model. Fibrin deposition and thrombosis, neutrophil recruitment, and fibrosis, are pathologic features seen in end-stage murine tuberculosis, and these are mirrored in the pathology induced in our model, and also by TDM-bearing oil droplets reported by Retzinger et al. (55). Thus, TDM may have an important contribution to the progressive pathology of terminal tuberculosis, and the successful transmission of infectious bacteria.
In summary, our results demonstrate that trehalose mycolates are the principal biologically active species of BCG peripheral cell wall lipids known to traffic within BCG-infected macrophages. Trehalose mycolates are potent inducers of the cytokines IL-1, IL-6, and TNF-, as well as of neutrophil recruitment, and recapitulate many of the pathologic manifestations of tuberculosis. Our results indicate that cellular response to trehalose mycolates is independent of TLR2 or TLR4, but requires the TLR/IL-1R adaptor molecule MyD88. Further studies are required to identify the cellular recognition of TDM and the role(s) played by the lipid through the course of disease.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
TLR2–/– mice were the kind gift of Dr. Emil Unanue (Washington University, St. Louis MO). MyD88–/– mice were generated by Dr. Shizuo Akira and provided by Dr. Sabine Ehrt (Cornell University, New York, NY).
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grant HL55936.
2 Address correspondence and reprint requests to Dr. Elizabeth R. Rhoades at the current address: Lovelace Respiratory Research Institute, 2425 Ridgecrest Drive SE, Albuquerque, NM 87108. E-mail address: brhoades{at}lrri.org
3 Abbreviations used in this paper: BCG, bacillus Calmette-Guérin; PIM, phosphatidylinositol mannoside; PIM2, phosphatidylinositol dimannoside; CL, cardiolipin; TMM, trehalose monomycolate; TDM, trehalose dimycolate; MycB, mycoside B; BMM, bone marrow-derived macrophage; WT, wild type; VEGF, vascular endothelial growth factor.
Received for publication July 15, 2004. Accepted for publication December 22, 2004.
References
Russell, D. G.. 2001. Mycobacterium tuberculosis: here today, and here tomorrow. Nat. Rev. Mol. Cell Biol. 2:569.
Yamagami, H., T. Matsumoto, N. Fujiwara, T. Arakawa, K. Kaneda, I. Yano, K. Kobayashi. 2001. Trehalose 6,6'-dimycolate (cord factor) of Mycobacterium tuberculosis induces foreign-body- and hypersensitivity-type granulomas in mice. Infect. Immun. 69:810.
Silva, C. L.. 1985. Inflammation induced by mycolic acid-containing glycolipids of Mycobacterium bovis (BCG). Braz. J. Med. Biol. Res. 18:327.
Beatty, W. L., E. R. Rhoades, H. J. Ullrich, D. Chatterjee, J. E. Heuser, D. G. Russell. 2000. Trafficking and release of mycobacterial lipids from infected macrophages. Traffic 1:235.
Rhoades, E., F. Hsu, J. B. Torrelles, J. Turk, D. Chatterjee, D. G. Russell. 2003. Identification and macrophage-activating activity of glycolipids released from intracellular Mycobacterium bovis BCG. Mol. Microbiol. 48:875.
Brennan, P. J., H. Nikaido. 1995. The envelope of mycobacteria. Annu. Rev. Biochem. 64:29.
Briken, V., S. A. Porcelli, G. S. Besra, L. Kremer. 2004. Mycobacterial lipoarabinomannan and related lipoglycans: from biogenesis to modulation of the immune response. Mol. Microbiol. 53:391.
Billiau, A., P. Matthys. 2001. Modes of action of Freund’s adjuvants in experimental models of autoimmune diseases. J. Leukocyte Biol. 70:849.
Gilleron, M., C. Ronet, M. Mempel, B. Monsarrat, G. Gachelin, G. Puzo. 2001. Acylation state of the phosphatidylinositol mannosides from Mycobacterium bovis bacillus Calmette Guerin and ability to induce granuloma and recruit natural killer T cells. J. Biol. Chem. 276:34896.
Bekierkunst, A., I. S. Levij, E. Yarkoni, E. Vilkas, A. Adam, E. Lederer. 1969. Granuloma formation induced in mice by chemically defined mycobacterial fractions. J. Bacteriol. 100:95.
Lima, V. M., V. L. Bonato, K. M. Lima, S. A. Dos Santos, R. R. Dos Santos, E. D. Goncalves, L. H. Faccioli, I. T. Brandao, J. M. Rodrigues-Junior, C. L. Silva. 2001. Role of trehalose dimycolate in recruitment of cells and modulation of production of cytokines and NO in tuberculosis. Infect. Immun. 69:5305.
Medzhitov, R.. 2001. Toll-like receptors and innate immunity. Nat. Rev. Immunol. 1:135.
Abel, B., N. Thieblemont, V. J. Quesniaux, N. Brown, J. Mpagi, K. Miyake, F. Bihl, B. Ryffel. 2002. Toll-like receptor 4 expression is required to control chronic Mycobacterium tuberculosis infection in mice. J. Immunol. 169:3155.
Jones, B. W., T. K. Means, K. A. Heldwein, M. A. Keen, P. J. Hill, J. T. Belisle, M. J. Fenton. 2001. Different Toll-like receptor agonists induce distinct macrophage responses. J. Leukocyte Biol. 69:1036.
Rhoades, E. R., R. E. Geisel, B. A. Butcher, S. P. McDonough, and D. G. Russell. 2005. Cell wall lipids from Mycobacterium bovis BCG are inflammatory when inoculated within a gel matrix: characterization of a new model of the granulomatous response to mycobacterial components. Tuberculosis (Edinb). In press.
Perez, R. L., J. Roman, S. Roser, C. Little, M. Olsen, J. Indrigo, R. L. Hunter, J. K. Actor. 2000. Cytokine message and protein expression during lung granuloma formation and resolution induced by the mycobacterial cord factor trehalose-6,6'-dimycolate. J. Interferon Cytokine Res. 20:795.
VanHeyningen, T. K., H. L. Collins, D. G. Russell. 1997. IL-6 produced by macrophages infected with Mycobacterium species suppresses T cell responses. J. Immunol. 158:330.
Nagabhushanam, V., A. Solache, L. Ting, C. J. Escaron, J. Zhang, J. D. Ernst. 2003. Innate inhibition of adaptive immunity: Mycobacterium tuberculosis-induced IL-6 inhibits macrophage responses to IFN-. J. Immunol. 171:4750.
Sakaguchi, I., N. Ikeda, M. Nakayama, Y. Kato, I. Yano, K. Kaneda. 2000. Trehalose 6,6'-dimycolate (cord factor) enhances neovascularization through vascular endothelial growth factor production by neutrophils and macrophages. Infect. Immun. 68:2043.
Retzinger, G. S., S. C. Meredith, R. L. Hunter, K. Takayama, F. J. Kezdy. 1982. Identification of the physiologically active state of the mycobacterial glycolipid trehalose 6,6'-dimycolate and the role of fibrinogen in the biologic activities of trehalose 6,6'-dimycolate monolayers. J. Immunol. 129:735.
Seiler, P., P. Aichele, S. Bandermann, A. E. Hauser, B. Lu, N. P. Gerard, C. Gerard, S. Ehlers, H. J. Mollenkopf, S. H. Kaufmann. 2003. Early granuloma formation after aerosol Mycobacterium tuberculosis infection is regulated by neutrophils via CXCR3-signaling chemokines. Eur. J. Immunol. 33:2676.
Fulton, S. A., S. M. Reba, T. D. Martin, W. H. Boom. 2002. Neutrophil-mediated mycobacteriocidal immunity in the lung during Mycobacterium bovis BCG infection in C57BL/6 mice. Infect. Immun. 70:5322.
Pedrosa, J., B. M. Saunders, R. Appelberg, I. M. Orme, M. T. Silva, A. M. Cooper. 2000. Neutrophils play a protective nonphagocytic role in systemic Mycobacterium tuberculosis infection of mice. Infect. Immun. 68:577.
Appelberg, R., A. G. Castro, S. Gomes, J. Pedrosa, M. T. Silva. 1995. Susceptibility of beige mice to Mycobacterium avium: role of neutrophils. Infect. Immun. 63:3381.
Cardona, P. J., R. Llatjos, S. Gordillo, J. Diaz, B. Vinado, A. Ariza, V. Ausina. 2001. Towards a ‘human-like’ model of tuberculosis: intranasal inoculation of LPS induces intragranulomatous lung necrosis in mice infected aerogenically with Mycobacterium tuberculosis. Scand. J. Immunol. 53:65.
Kipnis, A., R. J. Basaraba, J. Turner, I. M. Orme. 2003. Increased neutrophil influx but no impairment of protective immunity to tuberculosis in mice lacking the CD44 molecule. J. Leukocyte Biol. 74:992.
Sugawara, I., H. Yamada, C. Li, S. Mizuno, O. Takeuchi, S. Akira. 2003. Mycobacterial infection in TLR2 and TLR6 knockout mice. Microbiol. Immunol. 47:327.
Sugawara, I., H. Yamada, S. Hua, S. Mizuno. 2001. Role of interleukin (IL)-1 type 1 receptor in mycobacterial infection. Microbiol. Immunol. 45:743.
Sugawara, I., H. Yamada, S. Mizuno, K. Takeda, S. Akira. 2003. Mycobacterial infection in MyD88-deficient mice. Microbiol. Immunol. 47:841.
D’Souza, C. D., A. M. Cooper, A. A. Frank, R. J. Mazzaccaro, B. R. Bloom, I. M. Orme. 1997. An anti-inflammatory role for T lymphocytes in acquired immunity to Mycobacterium tuberculosis. J. Immunol. 158:1217.
Scanga, C. A., A. Bafica, C. G. Feng, A. W. Cheever, S. Hieny, A. Sher. 2001. MyD88-deficient mice display a profound loss in resistance to Mycobacterium tuberculosis associated with partially impaired Th1 cytokine and nitric oxide synthase 2 expression. Infect. Immun. 72:2400.
Benelli, R., A. Albini, D. Noonan. 2003. Neutrophils and angiogenesis: potential initiators of the angiogenic cascade. Chem. Immunol. Allergy 83:167.
Saita, N., N. Fujiwara, I. Yano, K. Soejima, K. Kobayashi. 2000. Trehalose 6,6'-dimycolate (cord factor) of Mycobacterium tuberculosis induces corneal angiogenesis in rats. Infect. Immun. 68:5991.
Dannenberg, A. M., Jr, F. M. Collins. 2001. Progressive pulmonary tuberculosis is not due to increasing numbers of viable bacilli in rabbits, mice and guinea pigs, but is due to a continuous host response to mycobacterial products. Tuberculosis (Edinb) 81:229.
Turner, O., R. Basaraba, A. A. Frank, I. M. Orme. 2003. Granuloma formation in mouse and guinea pig models of experimental tuberculosis. D. L. Boros, Jr, ed. Granulomatous Infections and Inflammations: Cellular and Molecular Mechanisms 65. ASM Press, Washington, D.C.
Rhoades, E. R., A. A. Frank, I. M. Orme. 1997. Progression of chronic pulmonary tuberculosis in mice aerogenically infected with virulent Mycobacterium tuberculosis. Tuber. Lung Dis. 78:57.
Orme, I. M.. 1998. The immunopathogenesis of tuberculosis: a new working hypothesis. Trends Microbiol. 6:94.
Retzinger, G. S., S. C. Meredith, K. Takayama, R. L. Hunter, F. J. Kezdy. 1981. The role of surface in the biological activities of trehalose 6,6'-dimycolate: surface properties and development of a model system. J. Biol. Chem. 256:8208.(Rachel E. Geisel, Kaori S)
The hallmark of Mycobacterium-induced pathology is granulomatous inflammation at the site of infection. Mycobacterial lipids are potent immunomodulators that contribute to the granulomatous response and are released in appreciable quantities by intracellular bacilli. Previously we investigated the granulomagenic nature of the peripheral cell wall lipids of Mycobacterium bovis bacillus Calmette-Guérin (BCG) by coating the lipids onto 90-μm diameter microspheres that were mixed into Matrigel matrix with syngeneic bone marrow-derived macrophages and injected i.p. into mice. These studies demonstrated that BCG lipids elicit proinflammatory cytokines and recruit leukocytes. In the current study we determined the lipids responsible for this proinflammatory effect. BCG-derived cell wall lipids were fractionated and purified by liquid chromatography and preparative TLC. The isolated fractions including phosphatidylinositol dimannosides, cardiolipin, phosphatidylglycerol, phosphatidylethanolamine, trehalose monomycolate, trehalose dimycolate, and mycoside B. Trehalose dimycolate, when delivered to bone marrow-derived murine macrophages, induced the greatest secretion of IL-1, IL-6, and TNF- in vitro. Trehalose dimycolate similarly induced the greatest secretion of these proinflammatory cytokines in ex vivo matrices over the course of 12 days. Trehalose monomycolate and dimycolate also induced profound neutrophil recruitment in vivo. Experiments with TLR2 or TLR4 gene-deficient mice revealed no defects in responses to trehalose mycolates, although MyD88-deficient mice manifested significantly reduced cell recruitment and cytokine production. These results demonstrate that the trehalose mycolates, particularly trehalose dimycolate, are the most bioactive lipids in the BCG extract, inducing a proinflammatory cascade that influences granuloma formation.
Introduction
Despite a worldwide vaccination program and half a century of treatment with antibiotics, Mycobacterium tuberculosis, the causative agent of tuberculosis, remains one of the leading causes of death attributable to an infectious agent. The hallmark of infection with pathogenic Mycobacterium spp. is the formation of a granuloma. Classical tuberculoid granulomas consist of a nidus of infected phagocytes, surrounded by a layer of activated giant cells and epitheloid macrophages, with a mantle of lymphocytes and fibrosis at the periphery (1). The tuberculoid granuloma is a fascinating example of a balanced interplay between host and pathogen. The granuloma is essential for containment of tuberculosis, as activated inflammatory cells enclose the tubercle bacilli, preventing dissemination and transmission of the pathogen. However the structure of the granuloma sequesters viable bacilli in an environment distant from macrophage-activating T cells, thus appearing to protect the bacilli from elimination. The tuberculoid granuloma is also continually evolving, and the development of extensive necrosis and caseation in advanced disease allows for transmission of the bacteria. Hence, although the host may require the granuloma for containment, it is advantageous for the bacterium to modulate this structure to ensure both persistence, and eventually transmission.
Mycobacterial glycolipids have long been recognized to have immunomodulatory activity, notably the induction of granulomatous responses. Mycobacterial cell wall glycolipids induce foreign body-type granulomas in mice (2, 3), and we have shown that glycolipids of Mycobacterium bovis bacillus Calmette-Guérin (BCG)3 traffic within the host macrophage and spread to uninfected bystander macrophages (4). Released glycolipids included phosphatidylinositol dimannosides (PIM2), phosphatidylglycerol and cardiolipin (CL), phosphatidylethanolamine, trehalose monomycolate (TMM), trehalose dimycolate (TDM), and mycoside B (MycB) (5). These classes of lipids are shared with M. tuberculosis, and lipids from both species exert potent adjuvant-like effects. The mechanisms by which the mycobacterial lipids exert these effects have yet to be determined (5, 6, 7, 8).
A considerable body of data has been published detailing the in vivo activity of mycobacterial lipids. PIM2 and phosphatidylinositol hexamannoside are granulomagenic when delivered with alum, recruiting macrophages, neutrophils, and NKT cells. These properties are dependent upon the lipid portion of the molecule (9, 10). PIM2, when presented on 1-μm diameter microspheres, elicits TNF-, reactive nitrogen intermediates, and macrophage chemotactic protein-1 from IFN--primed bone marrow-derived macrophages (BMM) or thioglycolate-elicited peritoneal macrophages (5). TDM, also known as cord factor, can reproduce the pathophysiologic aspects of M. tuberculosis infection including granuloma formation and induction of proinflammatory cytokines. Doses of TDM as little as 1–5 μg injected i.v. are granulomagenic in the lungs of mice, and repeated doses are more pathogenic than a single, larger dose (11). TDM-associated pathology includes weight loss due to TNF--mediated cachexia, with associated hyperlipidemia, hypoglycemia and peritonitis with ischemic and hemorrhagic lesions in the gastrointestinal tract (12). TDM-coated biodegradable microspheres, when embolized into murine lungs, induce intense inflammatory infiltrates and IL-4, IL-6, TNF-, IL-12, IL-10, and IFN- (13). Hence, a variety of mycobacterial lipids have broad immunomodulatory capabilities that affect granuloma formation.
Microbial components or pathogen-associated molecular patterns are recognized predominantly by pattern-recognition receptors such as TLRs that play an important role in innate response to infection. Pattern-recognition receptors involved in recognizing M. tuberculosis include the mannose receptor, complement receptors, and TLRs (14). TLRs have an important role in triggering innate response to mycobacterial cell wall components, and the principal TLRs involved in recognition of mycobacterial lipids are TLR2 and TLR4. TLR2 binds mycobacterial lipoarabinomannan and phosphatidylinositol mannoside (PIM), and PIM also signals through TLR4 (14, 15, 16).
Although there are many studies detailing the bioactivity of individual mycobacterial cell wall components there has been no comparative analysis of the lipids known to be released within infected macrophages. The in vivo response of mice to the different lipid fractions was investigated using a lipid-coated bead/matrix model that was developed and described in an earlier study (17). In this model total BCG lipids induced proinflammatory cytokines and chemokines including TNF-, IL-1, IL-1, IL-6, macrophage chemotactic protein-1, IFN--inducible protein-10, and recruited polymorphonuclear and mononuclear leukocytes and lymphocytes. The current comparative study demonstrates that trehalose mycolates are the principal bioactive lipids of BCG when coated onto 90-μm diameter microspheres, eliciting cytokines in vitro and in vivo with associated recruitment of primarily neutrophils and macrophages. Intriguingly the mode of presentation of the lipids on large particles played a significant role in lipid bioactivity. We also show that the inflammatory response to TDM was not diminished in TLR2–/– or TLR4 defective mice, although cell recruitment and cytokine production in MyD88–/– mice was severely diminished indicating involvement of the TLR/IL-1R signaling pathway.
Materials and Methods
Mice
Age- and sex-matched wild-type (WT) mice including C57BL/6, C3H/HeN, and C57BL/6/129F2 strains were purchased from Charles River Breeding Laboratories and housed under specific pathogen-free conditions. Unless otherwise stated experiments were performed in C57BL/6 mice. Female 6- to 8-wk-old TLR4-deficient C3H/HeJ mice were purchased from The Jackson Laboratory. Male TLR2–/– mice were the kind gift of Dr. E. Unanue (Washington University, St. Louis, MO). MyD88-deficient mice on a C57BL/6/129F2 background were generated by Dr. S. Akira and provided by Dr. S. Ehrt (Cornell University, New York, NY).
Macrophage culture
Mice were anesthetized and sacrificed by cervical dislocation, and the femurs and tibias were harvested. Bone marrow was flushed from the bones with cold DMEM (Invitrogen Life Technologies), supplemented with L-929 cell-conditioned medium, heat-inactivated FCS (10%; Summit Biotechnology) and horse serum (5%; Sigma-Aldrich), L-glutamine (2 mM), sodium pyruvate (1 mM), 100 U/ml penicillin, 100 μg/ml streptomycin (Invitrogen Life Technologies). Bone marrow cells were cultured in bacterial grade 90-mm2 petri dishes (Valmark; Kord) at 37°C, 5% CO2 for 5 days, after which BMM were harvested in endotoxin-free PBS (Invitrogen Life Technologies) or expanded for an additional 3–5 days. BMM were used for in vitro or in vivo assays.
Mycobacterial culture and extraction of lipids
M. bovis BCG (Pasteur strain) was grown to late-log phase in Middlebrook 7H9 medium (Difco) supplemented with oleic acid, BSA-Fraction V, dextrose, NaCl and 0.05% Tween 80. Bacilli were washed extensively in 0.05% Tween 80/PBS followed by a detergent-free wash before extraction in chloroform and methanol as previously described (5). Briefly, the pellet was extracted twice in chloroform/methanol (2:1 v/v) for 15 min at 55°C with sonication. Bacterial debris was removed by passing the extract through a 0.2-μm Teflon filter, and hydrophilic contaminants were removed using a Folch wash. The extract was fractionated by liquid chromatography over silica gel 60 (EM Science) in increasing amounts of methanol in chloroform followed by preparative TLC on aluminum-backed silica gel 60 plates (EM Science). Individual lipids were scraped from the plate and extracted from the silica in chloroform: methanol (2:1 v/v) and purified further using reverse-phase chromatography through C18 Sep-pak columns (Waters). Lipids were detected on TLC plates using 50% ethanolic sulfuric acid and charring. Purified lipid fractions included those that had been identified previously: diphosphatidylglycerol CL, phosphatidylglycerol, phosphatidylethanolamine, PIM2a and PIM2b (with PIM2b having an additional acyl group attached to a mannose residue), TMM, TDM, and MycB (5). Lipids were resuspended in chloroform/methanol (2:1 v/v) at 10 mg/ml and stored at –20°C under nitrogen.
In vitro macrophage assays
BMM were plated overnight in 24-well plates at 2 x 105 per well in L929-supplemented DMEM. Sterile polystyrene microspheres (90-μm diameter; Polysciences) were coated passively in an excess of total BCG lipid extract, purified lipid fractions, or bovine-derived phosphatidylglycerol (Sigma-Aldrich) in PBS using alternate cycles of vortexing and sonication at 55°C. Coated microspheres were washed and resuspended in endotoxin-free PBS at 2% solids. For lipid comparison in vitro, coated microspheres were added to macrophage monolayers at an equivalent ratio to that used in the in vivo model, and noncoated beads or nonpulsed macrophages were included as controls. For bead size assays, an equivalent surface area of microspheres (1, 2, 3, 4.5, 6, 10, 25, and 90 μm diameter) was coated with an excess of TDM as earlier described. Microspheres were added to macrophages prepared as previously discussed, over a dose value of 0.5x, 1x, and 2x with 1x equivalent to a total bead surface area of 5 x 107 μm2. Cells were incubated for 48 h at 37°C in 5% CO2, after which supernatants were collected and stored at –70°C until analysis.
Delivery and harvest of lipid-bearing matrices
Polystyrene microspheres (90 μm in diameter) were coated with lipid and mixed (6.7% solids) with syngeneic BMM (1 x 107/ml) in ice-cold, phenol red-free, growth factor-reduced Matrigel (BD Biosciences), a semidefined gel composed of murine laminin, collagen (type IV), entactin, and minimal concentrations of growth factors (epidermal growth factor, nerve growth factor, platelet-derived growth factor, and TGF-). Liquid matrix (300 μl) was injected i.p. into mice upon which it polymerized at body temperature. At various times postinjection mice were sacrificed and matrices were recovered and washed in endotoxin-free PBS (Invitrogen Life Technologies).
Histology
Matrices were recovered at various times postinjection and fixed in 4% neutral-buffered paraformaldehyde. Matrices were mounted in Histogel (Richard-Allan Scientific) before routine histologic processing and paraffin embedding. Sections (4-μm thickness) were stained with H&E. Representative micrographs were captured by an AxioCam on a Zeiss Axioskop II.
Cytokine ELISA
Peritoneal matrices were harvested at various times postinjection, and cultured ex vivo for an additional 48 h. Briefly, preweighed matrices (30 mg) were cultured in DMEM containing FCS, L-glutamine, sodium pyruvate, penicillin/streptomycin at 37°C 5% CO2. Culture supernatants were collected and stored at –70°C until analysis. TNF-, IL-1, IL-2, IL-4, IL-6, and IFN- were measured using BD OptEIA ELISA kits (BD Pharmingen). Vascular endothelial growth factor (VEGF) was measured using a Quantikine M ELISA kit (R&D Systems). Supernatants from individual mice (n = 3) were assayed and normalized to the weight of the matrix.
Flow cytometry
Peritoneal matrices were recovered, weighed, incubated briefly in cold BD Cell Recovery solution (BD Biosciences) and dispersed by passage through a 70-μm mesh. The numbers of viable cells were counted in trypan blue to calculate the number of viable cells per milligram of matrix. For immunophenotyping, cells were blocked in FACS buffer (ice-cold PBS with 10% heat-inactivated FCS and 5% normal mouse serum; Jackson ImmunoResearch Laboratories) and anti-FcRIII/RII (clone 2.4G2; BD Pharmingen) for 15 min. Cells (from individual mice (n = 3) or pooled samples) were subsequently stained in FACS buffer using fluorophore-conjugated Ab for 45 min at 4°C, and propidium iodide at 3 μg/ml (BD Pharmingen) was added during the final 10 min. Stained cells were washed in FACS buffer and resuspended in 10% FCS in PBS for immediate flow cytometric analyses.
Live leukocytes were gated upon as propidium iodide-negative events with a scatter profile typical of a mixed leukocyte population. Within this gate, neutrophils were identified as small Ly-6GhighF4/80– cells, macrophages were large Ly-6GvariableF4/80+I-A/I-E+ cells, and eosinophils were small, granular F4/80midI-A/I-E– events. B cells were agranular CD19+ events, T cells were small, CD3+ events, and NK cells were small, agranular CD49b+ (CD3+ or CD3–) events. Abs purchased from Caltag Laboratories included: anti-Ly-6G (clone RB6–8C5), anti-macrophage marker (clone F4/80), anti-CD49b (clone DX5), and anti-CD19 (clone 6D5). Abs purchased from BD Pharmingen included: anti-I-A/I-E (clone 2G9), anti-CD3 (clone 145-2C11), and isotype control Abs. Samples were analyzed on a FACSCalibur equipped with two lasers (488 and 633 nm) and analyzed using CellQuest software (BD Biosciences). Data were presented as the number of viable leukocytes for each subset per milligram of matrix (percentage of viable leukocytes x number of viable cells per milligram of gel).
Statistical methods
Cytokine measurements and calculations of cell numbers were performed on samples from individual mice and the SD of the mean values (n = 3) are presented. The statistical significance of differences in the means of samples relative to phosphatidylglycerol-induced responses was calculated using a Student’s paired t test (p < 0.05 and p < 0.005). Statistical differences in cell recruitment and cytokine production from TDM-bearing matrices of MyD88–/– and WT animals are indicated (p < 0.005).
Results
BCG lipid fractions elicit proinflammatory cytokines when delivered to BMM in vitro
Previous studies determined that total BCG lipids, when delivered i.p. into mice on 90-μm diameter beads with syngeneic BMM in the matrix, induced both leukocyte recruitment and proinflammatory cytokine production (17). To determine which of the lipids were proinflammatory, total BCG lipids were fractionated and the principal species were purified (Fig. 1). These lipids were coated onto 90-μm diameter beads and screened on BMM in vitro. Purified lipids included PIM2a (diacyl), PIM2b (triacyl), diphosphatidylglycerol CL, phosphatidylglycerol, phosphatidylethanolamine, TMM, TDM, and MycB. These had been identified previously as the major lipid species released from intracellular BCG and had been shown to traffic to neighboring, uninfected bystander macrophages (5).
FIGURE 1. Fractionated lipids of M. bovis BCG. Total peripheral cell wall lipids were extracted from cultures of BCG in 2:1 (v/v) chloroform/methanol. Lipids were fractionated by liquid chromatography and preparative TLC. A, Resolved in chloroform/methanol (90:10 v/v) to distinguish TMM, TDM, and MycB. B, Resolved in chloroform/methanol/water (65:25:4 v/v) to separate diacyl PIM2a, triacyl PIM2b, CL/diphosphatidylglycerol, and phosphatidylethanolamine (PE). Lipids are compared (A and B) with total BCG peripheral cell wall lipid extract (Total).
Secretion of IL-1, IL-6, and TNF- from BMM was elicited by some individual lipid fractions or by the total BCG lipid extract (Fig. 2). TDM was the principal inducer of IL-1, eliciting higher amounts than total BCG lipids (Fig. 2A). Several lipids induced IL-6 significantly above controls, including CL, TMM, TDM, PIM, and total BCG lipids (Fig. 2B). The trehalose mycolates and total BCG lipids were the strongest inducers of TNF-, although substantial amounts of the cytokine were detected from CL- and PIM-treated cells (Fig. 2C). PE or MycB failed to induce these cytokines. Consequently, CL, PIM2b, TMM, and TDM were selected for further analyses in vivo. PIM2b was selected for further investigation because it was slightly more stimulatory than PIM2a.
FIGURE 2. BCG-derived trehalose mycolates induce cytokines from BMM when delivered on 90-μm polystyrene microspheres. Lipid-coated polystyrene microspheres (90-μm diameter) were overlaid onto confluent monolayers of C57BL/6 BMM. Beads were added to macrophages at a ratio equivalent to that used for in vivo matrices. phosphatidylglycerol-coated beads (PG), noncoated beads (Bead), and nontreated macrophages () were included as controls. Cells were incubated for 48 h, and supernatants were analyzed by sandwich ELISA for IL-1 (A), IL-6 (B), and TNF- (C). Data from triplicate samples from a representative experiment are shown as mean ± SD, and the statistical significance is shown relative to phosphatidylglycerol (*, p < 0.05; **, p < 0.005).
Previous studies had determined that PIM2a or PIM2b, when presented on 1-μm diameter beads to IFN--primed BMM or thioglycolate-elicited peritoneal macrophages, was the principal proinflammatory lipid of the total BCG lipid extract (5). This finding was in contrast with the current study, which indicated that trehalose mycolates were the principal inflammatory lipids when presented on 90-μm diameter beads. To examine the effect of bead size on lipid bioactivity, the same batch of purified peripheral cell wall lipids of BCG were presented to unprimed or IFN--primed BMM on either 1-μm diameter or 90-μm diameter beads, and the secretion of TNF- was assayed. These studies recapitulated earlier results with PIM2 being the principal inducer of TNF- when presented on 1-μm diameter. beads (5). In contrast, trehalose mycolates elicited no TNF- on the smaller particles but were the most proinflammatory fraction when presented on 90-μm diameter beads to unprimed BMM (data not shown). To determine the bead diameter necessary for trehalose mycolate bioactivity, TDM was presented to unprimed BMM on microspheres of increasing size from 1- to 90-μm diameter, at a range of concentrations (Fig. 3). TNF- was elicited by beads of 10-μm diameter, which were phagocytosed; however cytokine secretion was maximal and equivalent for larger (25- or 90-μm diameter) beads to which the macrophages attached but did not internalize. These results indicate that the bioactivity of trehalose mycolates is highly dependent on the mechanism of presentation.
FIGURE 3. The bioactivity of TDM is dependent upon the size of the particle. Trehalose mycolate-coated polystyrene microspheres of increasing diameter (1, 2, 3, 4.5, 6, 10, 25, and 90 μm) were overlaid onto confluent monolayers of BMM. The numbers of beads delivered to the monolayers were normalized to bead surface area, and beads were added over a range with 1x being equivalent to a total bead surface area of 5 x 107 μm2/well. Cells were incubated for 48 h, and the supernatants were analyzed by ELISA for TNF-. Data from triplicate samples from a representative experiment are shown as mean ± SD.
Trehalose mycolates elicit proinflammatory cytokines in peritoneal matrices ex vivo
CL, TMM, TDM, and PIM2b were delivered to mice in matrices, and the inflammatory response in the recovered matrices was followed over 12 days. C57BL/6 mice were injected with matrices containing syngeneic BMM and 90-μm diameter beads coated with fractionated lipids, total lipids, or bovine phosphatidylglycerol. Matrices were harvested from mice after 14 h and after 4, 7, and 12 days, weighed, and cultured for an additional 48 h to compare levels of cytokine secretion. Minimal cytokine secretion had been measured from ex vivo cultured matrices that contained phosphatidylethanolamine, PIM2a, or MycB at 14 h and at 4 days, indicating that these lipids were not proinflammatory (data not shown). In contrast, TDM induced high levels of IL-1, IL-6, and TNF- (Fig. 4) with notably enhanced secretion of IL-1 that peaked at 14 h (Fig. 4A). TMM also induced IL-1, albeit at lower levels at 14 h, which declined after 4 days. Total BCG lipids induced moderate levels of IL-1 with similar kinetics to TDM. In contrast to the trehalose mycolate containing fractions, PIM2b and CL failed to elicit IL-1 at levels greater than the control lipid, phosphatidylglycerol. There was sustained secretion of IL-6 in response to TMM, TDM, and total BCG lipids (Fig. 4B). Levels of IL-6 induced by CL and PIM2b were similar to phosphatidylglycerol, increasing slowly over 12 days. TMM and TDM induced comparably high secretion of TNF- that peaked within 14 h (Fig. 4C). Previously, we had noted that BCG lipid-bearing matrices tended to attach to peritoneal tissue and vascularize (17). We investigated whether VEGF, a powerful proangiogenic factor was involved in the response to the bacterial lipids. Secretion of VEGF from TMM, TDM, and total BCG lipid matrices peaked at 14 h, but by 4 days secretion of VEGF was comparable in all matrices (Fig. 4D). There were no differences in IL-2, IL-4, or IFN- production stimulated by any of the lipid fractions, although low levels were detected at 7 days (data not shown). These data establish that trehalose mycolates are the major proinflammatory lipids in the total BCG extract.
FIGURE 4. BCG-derived trehalose mycolates also induce cytokines in peritoneal matrices. Mice were injected with matrices containing syngeneic C57BL/6 BMM, and the purified BCG lipids CL, PIM2b, TMM or TDM, total BCG lipids (Total), or bovine-derived phosphatidylglycerol (PG), and the matrices were recovered at the indicated times and cultured for an additional 48 h. Supernatants were collected and analyzed by ELISA. Levels of IL-1 (A), IL-6 (B), TNF- (C), and VEGF (D) were normalized to the weight of matrix. Data from triplicate samples from a representative experiment are shown as mean ± SD, and significantly different levels from phosphatidylglycerol-bearing matrices are indicated (*, p < 0.05; **, p < 0.005).
Trehalose mycolate-bearing matrices induce marked recruitment of cells, most notably of neutrophils
The recruitment of leukocytes to matrices bearing CL, PIM2b, TMM, or TDM was compared with those containing total BCG lipid or phosphatidylglycerol. BCG lipids recruited significantly more cells than phosphatidylglycerol, as shown previously (17); however, the trehalose mycolates recruited even greater numbers of viable cells through 12 days (Fig. 5A). TMM recruited more cells than total BCG lipids at 14 h and at 4 days before falling to similar levels as controls at 7 days. CL and PIM2b recruited similar numbers of cells to phosphatidylglycerol throughout the time course. To determine which leukocyte subsets were recruited to the matrices, cell suspensions were stained for flow cytometry. The principal leukocytes recruited to trehalose mycolates were neutrophils (Ly-6GhighF4/80–)(Fig. 5B). Neutrophil recruitment kinetics mirrored the kinetics of total cell numbers for the various lipid fractions. Macrophage (Ly-6GvariableF4/80+I-A/I-E+) numbers were steady throughout the 12-day time course with little difference between the fractions (Fig. 5C). TDM recruited significantly higher numbers of eosinophils (forward scatterlow, F4/80mid, I-A/I-E–) at 14 h (Fig. 5D). There were no significant differences in the numbers of NK cells (side scatterlow, CD49b+), T cells (CD3+), or B cells (CD19+) recruited to the various lipid fractions, and numbers were low relative to granulocytes (data not shown). These data indicate that the enhanced cell recruitment observed in trehalose mycolate-containing matrices is due predominantly to neutrophils.
FIGURE 5. Trehalose mycolates recruit enhanced numbers of neutrophils to peritoneal matrices. Mice were injected (i.p.) with matrices that contained syngeneic C57BL/6 BMM with BCG-derived CL, PIM2b, TMM, TDM, Total lipids, or bovine phosphatidylglycerol (PG). Single cell suspensions from preweighed matrices were harvested and counted in trypan blue to calculate the total number of viable cells per milligram of matrix (A). Cells were also stained for flow cytometry, excluding dead cells from the analysis using propidium iodide. The percentages of viable (propidium-iodide negative) leukocyte subsets were converted to total cell numbers. The number of neutrophils (Ly-6GhighF4/80–) (B), macrophages (Ly-6GvariableF4/80+I-A/I-E+) (C), and eosinophils (side scatterhighF4/80midI-A/I-E–) (D) are indicated. Data from triplicate samples from a representative experiment are shown as mean ± SD, and significantly different levels from phosphatidylglycerol-bearing matrices are indicated (*, p < 0.05; **, p < 0.005).
Gross and histologic features of BCG lipid fraction granulomas
To analyze tissue development in response to the various lipid fractions, matrices containing CL, PIM2b, TMM, TDM, total BCG lipid, or phosphatidylglycerol were recovered for histologic examination. In gross terms, PIM2b and CL matrices resembled phosphatidylglycerol matrices, consisting of small masses floating free within the peritoneal cavity (data not shown). Generally phosphatidylglycerol, PIM2b, and CL matrices were fractured into several individual pieces. Opacity and firmness of the masses increased with time reflecting the accumulation of cells and debris. Due to substantial cell recruitment, TMM and TDM matrices exhibited marked opacity and firmness at 14 h, equivalent to that seen in a 7-day total BCG matrix. Additionally, all of TDM and TMM matrices adhered to peritoneal tissue by 4 days postinjection. Matrices typically adhered to adipose tissue within the peritoneal cavity, usually to the broad ligament, omentum, and mesentery. Once adhered, the organization of the matrices rapidly changed with recruited cells migrating in from the adherence margin where accumulations of leukocytes were observed. Interestingly adherent matrices vascularized, exhibiting substantial vessel growth and hemorrhage. Trehalose mycolate matrices were more cohesive than all other matrices and usually formed one large mass within the peritoneum. Trehalose mycolate matrices, particularly TMM also induced hemorrhagic ascites, most markedly at 4 days postinjection (data not shown). Trehalose mycolate matrices were well vascularized and by 7 days large areas of hemorrhage were visible, particularly in the TMM matrices. In some animals receiving trehalose mycolate matrices, hyperemia and vasculature hypertrophy developed in the vessels of the peritoneal body wall. Splenomegaly and peritoneal lymphadenopathy were also observed and were maximal at 7 days (data not shown).
Fig. 6 displays photomicrographs of the lipid-containing matrices over a 12-day time course. CL, phosphatidylglycerol, and PIM2b (PIM2b shown in Fig. 6, A–D) matrices remained relatively acellular throughout the 12-day time course. These lipid fractions recruited a mixed leukocyte population, and cellular infiltrates were restricted to the margins of the matrix and to surrounding the coated beads. In comparison total BCG lipid matrices exhibited early neutrophil recruitment at 14 h (Fig. 6E) and peak lymphocyte recruitment at 7 days (Fig. 6G). Cell numbers within the total BCG lipid matrices were intermediate between the phosphatidylglycerol-like fractions and the trehalose mycolates. Histology confirmed that cell recruitment to matrices containing trehalose mycolates was more profound. By 14 h substantial neutrophil recruitment had occurred in matrices containing TMM (Fig. 6I) or TDM (Fig. 6M). In contrast to total BCG lipids, neutrophil recruitment to both TMM (Fig. 6J) and TDM (Fig. 6N) was sustained. By 7 days, spaces were evident in the TMM and TDM matrices, likely due to degradation of the matrix by the recruited cells; this was particularly evident in the vicinity of the beads (Fig. 6, K and O). In response to all lipid fractions, multinucleated giant cells were present peripheral to the beads by 12 days (indicated by arrows, Fig. 6, D and P).
FIGURE 6. Histology of lipid-bearing matrices. Matrices containing lipid-coated beads and syngeneic C57BL/6 BMM were injected into mice and harvested at 14 h and at 4, 7, or 12 days. Paraformaldehyde-fixed paraffin-embedded sections were stained with H&E. Photomicrographs (x200) were taken of the most cellular regions within each section. Circular holes present within the amorphous matrix occur where coated beads were lost during processing. Sections containing PIM2b (A–D) exhibited similar cellularity to phosphatidylglycerol (PG) and CL (data not shown). Sections with TMM (I–L) or TDM (M–P) were more cellular than those with total BCG lipid (E–H), exhibiting substantial neutrophil recruitment. Note that TMM promoted greater hemorrhage within the matrix than TDM, and this was noted both grossly and histologically.
Cellular response to trehalose mycolates is independent of TLR2 or TLR4 but dependent on MyD88
Responses to most pathogen-associated molecular patterns characterized to date signal through the TLRs, and TLR2 and TLR4 bind mycobacterial constituents. To determine whether the cellular response to TDM was dependent on TLR2 or TLR4, cytokine production and cell recruitment to TDM-bearing matrices was assessed in mice deficient in these receptors. TDM-bearing matrices with syngeneic BMM were injected i.p. into TLR4- or TLR2-deficient mice or WT controls, and the responses were compared against phosphatidylglycerol-bearing matrices. The matrices were assessed for cell recruitment and cytokine production at 14 h postinjection. Minimal cell recruitment was seen in the phosphatidylglycerol matrices in any of the mice; in contrast, there were substantial and comparable levels of cell recruitment to TDM (Fig. 7, A and E) in TLR-deficient and WT mice. Similarly there was no defect in the production of IL-1, IL-6, or TNF- in either TLR4-deficient (Fig. 7, B–D) or TLR2-deficient (Fig. 7, F–H) mice, and there were no histologic differences compared with WT mice (data not shown).
FIGURE 7. Responses to TDM by TLR2-, TLR4-, and MyD88-deficient mice. Matrices containing either TDM or phosphatidylglycerol (PG), and syngeneic BMM were injected (i.p.) into TLR4-, TLR2- or MyD88-deficient mice or WT controls (MyD88–/– phosphatidylglycerol group was not determined (N.D.) due to scarcity of gene-deletion animals). Matrices were recovered 14 h postinjection and analyzed for cell recruitment and cytokine production as previously described. TLR4-deficient mice exhibited no defects in response to TDM in terms of cell recruitment (A), or cytokine production (B–D) relative to phosphatidylglycerol controls. Likewise, TLR2-deficient mice exhibited no defect in cell recruitment (E) or cytokine production (F–H) to TDM, relative to controls. In contrast MyD88-deficient mice had profound defects in cell recruitment (I), and cytokine production (J–L) compared with WT. Data from a representative experiment are presented as mean of matrices from individual mice ± SD. Statistical significance is shown relative to phosphatidylglycerol in WT for each group (*, p < 0.05; **, p < 0.005). Statistical differences in cell recruitment and cytokine production from TDM-bearing matrices of MyD88–/– and WT animals is indicated (, p < 0.005). There were no statistically significant differences between TLR-deficient and WT animals.
As a preliminary determination of whether other TLRs could be implicated in the recognition of TDM, we examined mice deficient in the common TLR-signaling adapter MyD88. TDM-bearing matrices from MyD88–/– mice were assayed for cell recruitment (Fig. 7I) and secretion of IL-1, IL-6, or TNF- (Fig. 7, J–L). MyD88–/– mice exhibited a profound defect in cell recruitment and cytokine production in TDM-bearing matrices compared with WT animals. These results indicate that the response to TDM, although independent of TLR2 or TLR4, is attenuated by the absence of the TLR/IL-1 adaptor MyD88.
Discussion
This study compared the relative biological activities of individual cell wall lipids that are released by M. bovis BCG into host macrophages. Comparison of seven principal peripheral cell wall lipids showed that trehalose mycolates were the most potent proinflammatory agents. When coated onto 90-μm beads and incubated with macrophages in vitro or inoculated into mice within matrices, trehalose mycolates induced substantial levels of IL-1, IL-6, TNF-, and VEGF with marked recruitment of neutrophils in vivo.
The inflammatory and toxic natures of trehalose mycolates have been appreciated for half a century. The physiologic dose of TDM from pulmonary mycobacterial infection in mice has been estimated to be 5 μg, and mice respond to doses of 1–5 μg (11). In our study, the beads were coated in excess of TDM, delivering an estimated 3–4 μg to each mouse. Mycobacterium smegmatis-derived TDM, delivered as an oil-in-water emulsion, induces IL-1, IL-6, and TNF- expression in a lung granuloma model with no detectable changes in mRNA for IFN-, IL-2, IL-4, IL-10, or IL-12 (18). Similarly, in our model, trehalose mycolates did not induce IL-12p70, the TH1 cytokine, IFN-, or the TH2 cytokines IL-4 or IL-10. Cytokines induced by trehalose mycolates contribute to several aspects of tuberculosis. Levels of TNF- induced by infection or TDM correlate with inflammation at the tissue level (18). TNF- induced by TDM after repeated injections also induces cachexia, a prominent clinical feature in end-stage tuberculosis (12). The weight of mice was not monitored in our study although none exhibited any sign of physical distress. A striking feature of trehalose mycolates matrices was sustained induction of high levels of IL-6. BCG-infected BMM also release IL-6, which inhibits proliferation of 3A9 T cells on hen-egg lysozyme-pulsed APCs (19). M. tuberculosis-infected macrophages secreted IL-6,which inhibited the response of uninfected macrophages to IFN- (20). Similarly, the high levels IL-6 induced by trehalose mycolates may suppress T cell and macrophage responses in the microenvironment of the granuloma. Total BCG lipids, TMM or TDM induced similar levels of IL-6 suggestive of synergistic induction by total BCG lipids because the total concentration of TMM and TDM combined in total BCG lipids was less than in matrices containing the purified fractions.
The striking features of trehalose mycolate-bearing matrices were prominent neutrophil recruitment accompanied by adherence and neovascularization. Neutrophil recruitment has been reported in other models of TDM-elicited inflammation (21, 22). Silva (3) has described cell recruitment to BCG or charcoal particles that were coated with TDM or TMM and injected (i.v.), subsequently lodging in the lungs. In contrast to our study, a moderate mononuclear infiltrate was noted, whereas intense neutrophil infiltrates were the response to glycerol monomycolate and glucose monomycolate. Macrophage recruitment to the matrices remained constant irrespective of the nature of the lipid.
The role of neutrophils in immunity to tuberculosis remains controversial. Neutrophil depletion during early mycobacterial infection exacerbates disease and dysregulates granuloma formation (23, 24, 25, 26). However neutrophils are associated with tissue damage and necrosis in late disease and are a feature of granulomas in gene-deletion models associated with increased susceptibility to tuberculosis (27, 28, 29, 30, 31, 32, 33). Neutrophils release mediators that likely contributed to the angiogenesis in the trehalose mycolate-bearing matrices, evident at 4 days postinjection (34). Several leukocyte products play a role in angiogenesis, including VEGF, IL-1, IL-6, IL-8, basic fibroblast growth factor, reactive nitrogen intermediates, TNF-, and TGF- (35, 36, 37, 38, 39). An appreciation for the role of neutrophils in angiogenesis is growing; neutrophils are sources of proangiogenic mediators including VEGF, IL-8, hepatocyte growth factor, and platelet-activating factor, and they can recruit additional leukocytes, like macrophages that contribute further to the proangiogenic cascade (34, 40). VEGF, an acute-phase proinflammatory response cytokine, increases vascular permeability and monocyte recruitment. Sakaguchi et al. (21) demonstrated TDM-enhanced neovascularization of murine air pouches with Rhodococcus-derived TDM, and neovascularization was attributed to the presence of VEGF-expressing neutrophils and macrophages. TDM from M. tuberculosis also induces neovascularization in a rat cornea model by cytokine-dependent mechanisms (35). Neutralizing Abs to VEGF or IL-8 substantially inhibited neovascularization, whereas anti-TNF- produced a partial inhibition; however, the cellular source of the angiogenic cytokines was not determined. In our model, trehalose mycolates or total BCG lipids induced an early burst of VEGF secretion from ex vivo matrices, coincident with the enhanced arrival of neutrophils (Fig. 4D), suggesting a role for both VEGF and neutrophils in the vascularization of adherent matrices. Although the granule protein eosinophil peroxidase has inhibitory activity toward M. tuberculosis H37Rv in vitro, the presence of eosinophils in pulmonary models of tuberculosis is considered detrimental as they are reflective of a harmful TH2 cytokine profile and contribute to tissue necrosis (41, 42, 43). Eosinophil recruitment was noted in response to trehalose mycolates, and likely reflects a lack of TH1-bias in the cytokine profile of matrices.
It has been reported that repeated injections (i.p.) of TDM-coated beads induces abdominal pathology very similar to that observed in TDM- or TMM-bearing matrices, including prominent peritoneal adhesions, hyperemic abdominal vessels, abdominal lymphadenopathy, splenomegaly, and a bloody ascites. Adhesions in the peritoneal cavity upon injection of TDM-coated beads initially consisted of neutrophils, and later of macrophages and fibroblasts (44). In our model, most lipid-bearing matrices floated freely in the peritoneal cavity, with the recruited cells coming from the surrounding peritoneal exudate. However, all of the TMM and TDM matrices were adherent by 4 days postinjection, a frequency higher than noted previously for matrices containing total BCG lipids (17). Comparison of adherent and free-floating total BCG lipid-bearing matrices from the same animal revealed that adherent matrices were more cellular, with infiltrates penetrating further into the matrix along the zone of adherence as well as in the regions of neovascularization. Adherence is significant in this model as it allowed recruitment of cells directly from the inflamed tissues, reflecting the mode of cell recruitment to tuberculoid pulmonary granulomas. Adherence of all trehalose mycolate matrices in the present study suggests that these lipids are responsible for this phenomenon.
The mechanism that promotes adherence of trehalose mycolate-bearing matrices is unknown. Retzinger et al. (22) determined that fibrinogen is a potent cofactor in TDM monolayers on coated beads, inducing more rapid and increased footpad swelling. There was thrombin-like procoagulant activity in the peritoneal cavities of mice injected with TDM-coated beads, and maximum biological activity occurred when TDM-coated beads were coated with blood plasma, which contains fibrinogen. It was postulated that thrombin-dependent conversion of fibrinogen to fibrin in situ was necessary for the biological activity of TDM monolayers. Any exposed hydrophobic residues are likely to be coated with fibrinogen in vivo, and conversion to fibrin and formation of adhesions may explain the adherence of matrices to tissues in our model. It is possible that this procoagulant activity enhances the mechanical adhesiveness of trehalose mycolate-bearing matrices, and we are currently investigating this in the matrix model.
Mechanisms of cellular activation by trehalose mycolates have yet to be determined. M. tuberculosis or its cell wall components can stimulate cells via TLRs (45, 46, 47). TLR2 and TLR4 are the principal TLRs reported to mediate recognition of M. tuberculosis, however neither TLR2- nor TLR4-deficient mice show a defect in cytokine production or leukocyte recruitment in response to TDM. The literature concerning the roles of TLR in tuberculosis infections is varied and contradictory (31, 33). TLR2 appears to play a minor role although TLR2-deficient mice succumb more readily to high dose challenge with M. tuberculosis (48). TLR4-deficient mice show a chronic neutrophilic pneumonia with reduced production of proinflammatory cytokines following challenge with M. bovis BCG (15, 49). In contrast, experiments with MyD88, the adaptor molecule required for signaling by TLRs and IL-1R, have in most instances revealed an increased susceptibility to mycobacterial infection (33). In the current study, MyD88–/– mice were defective in cell recruitment and cytokine production to trehalose mycolates compared with WT controls. This suggests that trehalose mycolates activate macrophages using a different TLR-mediated or IL-1 family receptor-mediated pathway than TLR2 or TLR4. Further analysis is required to determine the receptor(s) involved in the cellular recognition and response to TDM.
It is interesting to note that in a previous in vitro comparative study, we showed that PIMs were the most stimulatory lipids when presented on 1-μm beads to BMM, whereas the same numbers of trehalose mycolate-coated beads did not induce cytokines (5). This disparity was so striking that we compared responses to the same batch of lipid coated onto either 1- or 90-μm diameter beads (data not shown). Trehalose mycolates presented on 90-μm beads, but not 1-μm beads, induced inflammatory cytokines in vitro and in vivo. When a range of bead sizes was examined TDM-coated 10-μm diameter beads that could be phagocytosed elicited some TNF-, but secretion was maximal and equivalent for 25- and 90-μm diameter beads to which the macrophages attached. These data suggest that factors such as the number or receptors ligated, the duration of receptor/ligand interaction, or the formation of an adhesive plaque might modulate the cell response to TDM. These results recapitulate the elegant studies of Hunter and colleagues (22, 50) who documented that the biological activity of TDM was dependent on its presentation as a planar surface, in a monolayer, on a large particle, or in an oil/water interface. Given these structural requirements for biological activity it is pertinent to ask what relevance this might have in the course of disease. During the late stages of tuberculosis in guinea pigs, rabbits, and humans there is loss of granuloma containment, and breakdown of cavitary lesions, with uncontrolled extracellular mycobacterial replication, and the accumulation of cellular and lipid debris that is associated with neutrophil recruitment (51). Mice do not typically form the classical central caseous necrosis of pulmonary granulomas, however accumulation of lipid and cellular debris, neutrophil recruitment and fibrosis are common features of end-stage tuberculosis (52, 53, 54). The pathologic effects of trehalose mycolates may be predominant in the decompensating phase of tuberculosis, when uncontrolled extracellular bacillary growth and bacterial pellicle formation, with the accumulation of lipid debris present the lipid in a more planar configuration, similar to that in our model. Fibrin deposition and thrombosis, neutrophil recruitment, and fibrosis, are pathologic features seen in end-stage murine tuberculosis, and these are mirrored in the pathology induced in our model, and also by TDM-bearing oil droplets reported by Retzinger et al. (55). Thus, TDM may have an important contribution to the progressive pathology of terminal tuberculosis, and the successful transmission of infectious bacteria.
In summary, our results demonstrate that trehalose mycolates are the principal biologically active species of BCG peripheral cell wall lipids known to traffic within BCG-infected macrophages. Trehalose mycolates are potent inducers of the cytokines IL-1, IL-6, and TNF-, as well as of neutrophil recruitment, and recapitulate many of the pathologic manifestations of tuberculosis. Our results indicate that cellular response to trehalose mycolates is independent of TLR2 or TLR4, but requires the TLR/IL-1R adaptor molecule MyD88. Further studies are required to identify the cellular recognition of TDM and the role(s) played by the lipid through the course of disease.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
TLR2–/– mice were the kind gift of Dr. Emil Unanue (Washington University, St. Louis MO). MyD88–/– mice were generated by Dr. Shizuo Akira and provided by Dr. Sabine Ehrt (Cornell University, New York, NY).
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by National Institutes of Health Grant HL55936.
2 Address correspondence and reprint requests to Dr. Elizabeth R. Rhoades at the current address: Lovelace Respiratory Research Institute, 2425 Ridgecrest Drive SE, Albuquerque, NM 87108. E-mail address: brhoades{at}lrri.org
3 Abbreviations used in this paper: BCG, bacillus Calmette-Guérin; PIM, phosphatidylinositol mannoside; PIM2, phosphatidylinositol dimannoside; CL, cardiolipin; TMM, trehalose monomycolate; TDM, trehalose dimycolate; MycB, mycoside B; BMM, bone marrow-derived macrophage; WT, wild type; VEGF, vascular endothelial growth factor.
Received for publication July 15, 2004. Accepted for publication December 22, 2004.
References
Russell, D. G.. 2001. Mycobacterium tuberculosis: here today, and here tomorrow. Nat. Rev. Mol. Cell Biol. 2:569.
Yamagami, H., T. Matsumoto, N. Fujiwara, T. Arakawa, K. Kaneda, I. Yano, K. Kobayashi. 2001. Trehalose 6,6'-dimycolate (cord factor) of Mycobacterium tuberculosis induces foreign-body- and hypersensitivity-type granulomas in mice. Infect. Immun. 69:810.
Silva, C. L.. 1985. Inflammation induced by mycolic acid-containing glycolipids of Mycobacterium bovis (BCG). Braz. J. Med. Biol. Res. 18:327.
Beatty, W. L., E. R. Rhoades, H. J. Ullrich, D. Chatterjee, J. E. Heuser, D. G. Russell. 2000. Trafficking and release of mycobacterial lipids from infected macrophages. Traffic 1:235.
Rhoades, E., F. Hsu, J. B. Torrelles, J. Turk, D. Chatterjee, D. G. Russell. 2003. Identification and macrophage-activating activity of glycolipids released from intracellular Mycobacterium bovis BCG. Mol. Microbiol. 48:875.
Brennan, P. J., H. Nikaido. 1995. The envelope of mycobacteria. Annu. Rev. Biochem. 64:29.
Briken, V., S. A. Porcelli, G. S. Besra, L. Kremer. 2004. Mycobacterial lipoarabinomannan and related lipoglycans: from biogenesis to modulation of the immune response. Mol. Microbiol. 53:391.
Billiau, A., P. Matthys. 2001. Modes of action of Freund’s adjuvants in experimental models of autoimmune diseases. J. Leukocyte Biol. 70:849.
Gilleron, M., C. Ronet, M. Mempel, B. Monsarrat, G. Gachelin, G. Puzo. 2001. Acylation state of the phosphatidylinositol mannosides from Mycobacterium bovis bacillus Calmette Guerin and ability to induce granuloma and recruit natural killer T cells. J. Biol. Chem. 276:34896.
Bekierkunst, A., I. S. Levij, E. Yarkoni, E. Vilkas, A. Adam, E. Lederer. 1969. Granuloma formation induced in mice by chemically defined mycobacterial fractions. J. Bacteriol. 100:95.
Lima, V. M., V. L. Bonato, K. M. Lima, S. A. Dos Santos, R. R. Dos Santos, E. D. Goncalves, L. H. Faccioli, I. T. Brandao, J. M. Rodrigues-Junior, C. L. Silva. 2001. Role of trehalose dimycolate in recruitment of cells and modulation of production of cytokines and NO in tuberculosis. Infect. Immun. 69:5305.
Medzhitov, R.. 2001. Toll-like receptors and innate immunity. Nat. Rev. Immunol. 1:135.
Abel, B., N. Thieblemont, V. J. Quesniaux, N. Brown, J. Mpagi, K. Miyake, F. Bihl, B. Ryffel. 2002. Toll-like receptor 4 expression is required to control chronic Mycobacterium tuberculosis infection in mice. J. Immunol. 169:3155.
Jones, B. W., T. K. Means, K. A. Heldwein, M. A. Keen, P. J. Hill, J. T. Belisle, M. J. Fenton. 2001. Different Toll-like receptor agonists induce distinct macrophage responses. J. Leukocyte Biol. 69:1036.
Rhoades, E. R., R. E. Geisel, B. A. Butcher, S. P. McDonough, and D. G. Russell. 2005. Cell wall lipids from Mycobacterium bovis BCG are inflammatory when inoculated within a gel matrix: characterization of a new model of the granulomatous response to mycobacterial components. Tuberculosis (Edinb). In press.
Perez, R. L., J. Roman, S. Roser, C. Little, M. Olsen, J. Indrigo, R. L. Hunter, J. K. Actor. 2000. Cytokine message and protein expression during lung granuloma formation and resolution induced by the mycobacterial cord factor trehalose-6,6'-dimycolate. J. Interferon Cytokine Res. 20:795.
VanHeyningen, T. K., H. L. Collins, D. G. Russell. 1997. IL-6 produced by macrophages infected with Mycobacterium species suppresses T cell responses. J. Immunol. 158:330.
Nagabhushanam, V., A. Solache, L. Ting, C. J. Escaron, J. Zhang, J. D. Ernst. 2003. Innate inhibition of adaptive immunity: Mycobacterium tuberculosis-induced IL-6 inhibits macrophage responses to IFN-. J. Immunol. 171:4750.
Sakaguchi, I., N. Ikeda, M. Nakayama, Y. Kato, I. Yano, K. Kaneda. 2000. Trehalose 6,6'-dimycolate (cord factor) enhances neovascularization through vascular endothelial growth factor production by neutrophils and macrophages. Infect. Immun. 68:2043.
Retzinger, G. S., S. C. Meredith, R. L. Hunter, K. Takayama, F. J. Kezdy. 1982. Identification of the physiologically active state of the mycobacterial glycolipid trehalose 6,6'-dimycolate and the role of fibrinogen in the biologic activities of trehalose 6,6'-dimycolate monolayers. J. Immunol. 129:735.
Seiler, P., P. Aichele, S. Bandermann, A. E. Hauser, B. Lu, N. P. Gerard, C. Gerard, S. Ehlers, H. J. Mollenkopf, S. H. Kaufmann. 2003. Early granuloma formation after aerosol Mycobacterium tuberculosis infection is regulated by neutrophils via CXCR3-signaling chemokines. Eur. J. Immunol. 33:2676.
Fulton, S. A., S. M. Reba, T. D. Martin, W. H. Boom. 2002. Neutrophil-mediated mycobacteriocidal immunity in the lung during Mycobacterium bovis BCG infection in C57BL/6 mice. Infect. Immun. 70:5322.
Pedrosa, J., B. M. Saunders, R. Appelberg, I. M. Orme, M. T. Silva, A. M. Cooper. 2000. Neutrophils play a protective nonphagocytic role in systemic Mycobacterium tuberculosis infection of mice. Infect. Immun. 68:577.
Appelberg, R., A. G. Castro, S. Gomes, J. Pedrosa, M. T. Silva. 1995. Susceptibility of beige mice to Mycobacterium avium: role of neutrophils. Infect. Immun. 63:3381.
Cardona, P. J., R. Llatjos, S. Gordillo, J. Diaz, B. Vinado, A. Ariza, V. Ausina. 2001. Towards a ‘human-like’ model of tuberculosis: intranasal inoculation of LPS induces intragranulomatous lung necrosis in mice infected aerogenically with Mycobacterium tuberculosis. Scand. J. Immunol. 53:65.
Kipnis, A., R. J. Basaraba, J. Turner, I. M. Orme. 2003. Increased neutrophil influx but no impairment of protective immunity to tuberculosis in mice lacking the CD44 molecule. J. Leukocyte Biol. 74:992.
Sugawara, I., H. Yamada, C. Li, S. Mizuno, O. Takeuchi, S. Akira. 2003. Mycobacterial infection in TLR2 and TLR6 knockout mice. Microbiol. Immunol. 47:327.
Sugawara, I., H. Yamada, S. Hua, S. Mizuno. 2001. Role of interleukin (IL)-1 type 1 receptor in mycobacterial infection. Microbiol. Immunol. 45:743.
Sugawara, I., H. Yamada, S. Mizuno, K. Takeda, S. Akira. 2003. Mycobacterial infection in MyD88-deficient mice. Microbiol. Immunol. 47:841.
D’Souza, C. D., A. M. Cooper, A. A. Frank, R. J. Mazzaccaro, B. R. Bloom, I. M. Orme. 1997. An anti-inflammatory role for T lymphocytes in acquired immunity to Mycobacterium tuberculosis. J. Immunol. 158:1217.
Scanga, C. A., A. Bafica, C. G. Feng, A. W. Cheever, S. Hieny, A. Sher. 2001. MyD88-deficient mice display a profound loss in resistance to Mycobacterium tuberculosis associated with partially impaired Th1 cytokine and nitric oxide synthase 2 expression. Infect. Immun. 72:2400.
Benelli, R., A. Albini, D. Noonan. 2003. Neutrophils and angiogenesis: potential initiators of the angiogenic cascade. Chem. Immunol. Allergy 83:167.
Saita, N., N. Fujiwara, I. Yano, K. Soejima, K. Kobayashi. 2000. Trehalose 6,6'-dimycolate (cord factor) of Mycobacterium tuberculosis induces corneal angiogenesis in rats. Infect. Immun. 68:5991.
Dannenberg, A. M., Jr, F. M. Collins. 2001. Progressive pulmonary tuberculosis is not due to increasing numbers of viable bacilli in rabbits, mice and guinea pigs, but is due to a continuous host response to mycobacterial products. Tuberculosis (Edinb) 81:229.
Turner, O., R. Basaraba, A. A. Frank, I. M. Orme. 2003. Granuloma formation in mouse and guinea pig models of experimental tuberculosis. D. L. Boros, Jr, ed. Granulomatous Infections and Inflammations: Cellular and Molecular Mechanisms 65. ASM Press, Washington, D.C.
Rhoades, E. R., A. A. Frank, I. M. Orme. 1997. Progression of chronic pulmonary tuberculosis in mice aerogenically infected with virulent Mycobacterium tuberculosis. Tuber. Lung Dis. 78:57.
Orme, I. M.. 1998. The immunopathogenesis of tuberculosis: a new working hypothesis. Trends Microbiol. 6:94.
Retzinger, G. S., S. C. Meredith, K. Takayama, R. L. Hunter, F. J. Kezdy. 1981. The role of surface in the biological activities of trehalose 6,6'-dimycolate: surface properties and development of a model system. J. Biol. Chem. 256:8208.(Rachel E. Geisel, Kaori S)