Published online before print July 5, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Biomedical Sciences, Florida Atlantic University, Boca Raton; and
Department of Physiology, Brody School of Medicine at East Carolina University, Greenville, North Carolina; and
Caswell Beach, North Carolina
1 Correspondence: Department of Biomedical Sciences, Florida Atlantic University, 777 Glades Rd., P.O. Box 3091, Boca Raton, FL 33431-0991. E-mail yshibata{at}fau.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, express COX-1 and COX-2 within 1 day but release only minimal amounts of PGE2 following elicitation with calcium ionophore A23187. For further characterization of in vivo requirements for development of PGE2-releasing MØ (PGE2-MØ), C57Bl/6 [wild-type (WT)], and interleukin (IL)-10-deficient (IL-10–/–) mice were treated intraperitoneally with heat-killed Mycobacterium bovis bacillus Calmette-Guerin (HK-BCG). One day following injection, COX-2 was induced in splenic MØ of both mouse strains. However, PGE2 biosynthesis by these MØ was not increased. Thus, expression of COX-2 is not sufficient to induce PGE2 production in vivo or in vitro. In sharp contrast, 14 days after HK-BCG treatment, PGE2 release by COX-2+ splenic MØ increased as much as sevenfold, and a greater increase was seen in IL-10–/– cells than in WT cells. To further determine whether the 14-day splenic PGE2-MØ could be derived from bone marrow precursors, we established a chimera in which bone marrow cells were transfused from green fluorescent protein (GFP)-transgenic donors to WT mice. Donors and recipients were treated with HK-BCG simultaneously, and marrow transfusion was performed on Days 1 and 2. On Day 14 after BCG treatment, a significant number of spleen cells coexpressed COX-2 and GFP, indicating that bone marrow-derived COX-2+ MØ may be responsible for the increased PGE2 production.
Key Words: PGE2 • releasing macrophage • F4/80+ • GFP • chimera
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In the production of PGE2, MØ metabolize endogenous arachidonic acid (AA) through cyclooxygenases [COX; PGH synthase (PGHS), EC 1.4.99.1], rate-limiting enzymes for PGH2 synthesis. PGHS-1 (COX-1) is a constitutive isoform, whereas PGHS-2 (COX-2) is induced in response to various inflammatory mediators [15 ]. PGH2 is converted to PGE2 by cytosolic PGES (cPGES) and microsomal PGES-1 (mPGES-1) [16 , 17 ]. PGE2-releasing MØ (PGE2-MØ) are ontogenically heterogeneous populations, and PGE2 biosynthesis appears to be associated with differential expression of isoforms of the key enzymes [9 , 18 ]. Although splenic COX-2+ MØ develop in response to various inflammatory conditions, the exact mechanisms for splenic PGE2-MØ formation appear to be complex and controversial [19 , 20 ]. Our in vitro studies suggested that unlike other tissue MØ, splenic MØ in inflammation may not release PGE2 in proportion to increased COX-2 expression [21 ]. Furthermore, endogenous IL-10, an immunosuppressive cytokine reported to inhibit COX-2 expression in bone marrow MØ, peritoneal MØ, neutrophils, and cancer cells [22 , 23 ], did not significantly modify COX-2 expression by these in vitro splenic MØ [21 ].
Previous murine studies [5
, 9
, 18
, 24
] indicate that intraperitoneal (i.p.) administration of HK-Mycobacterium bovis bacillus Calmette-Guerin (HK-BCG) or Corynebacterium parvum(Propionibacterium acnes) results in formation of splenic COX-2+ MØ and five- to tenfold increases in PGE2 release within 5–21 days. In previous whole animal studies, we found that the formation of splenic PGE2-MØ is dependent on radio-sensitive bone marrow precursors but is independent of circulating monocytes [9
, 18
]. The formation of PGE2-MØ appears to be down-regulated by endogenous IL-10 [24
]. Preliminary studies showed that 1 day after HK-BCG treatment, wild-type (WT) and IL-10–/–splenic MØ expressed COX-2 at relatively high levels in an IL-10-independent manner with little increase in the release of PGE2. This was consistent with our previous report that WT splenic MØ treated in vitro with lipopolysaccharide (LPS) and interferon- (IFN-) expressed COX-2 but did not produce substantially increased amounts of PGE2 [18
]. In the present study, we have further characterized differences in Day 1 COX-2+ MØ and Day 14 PGE2-MØ induced in vivo by HK-BCG treatment of IL-10–/– and WT mice.
An alternative explanation for our finding that radio-sensitive bone marrow supplies precursors of PGE2-MØ following treatment with HK-BCG is that an inflammatory cytokine "milieu" might convert resident splenic MØ to PGE2-MØ [19
, 20
]. However, the in vitro studies showed that PGE2-MØ do not develop from WT splenic MØ treated with the inflammatory mediators LPS and IFN- [21
]. Therefore, we hypothesized that the PGE2-MØ, present in the spleen on Day 14, are derived from bone marrow precursor cells and do not result from maturation of resident splenic cells. To determine more directly whether bone marrow precursor cells could populate the spleen following treatment with BCG, we have performed adoptive transfer of bone marrow from green fluorescent protein (GFP)-expressing mice. Adoptive transfer of donor bone marrow cells is widely used in research about the function and metabolism of lymphoid and myeloid cells. The transfer of cells from GFP-transgenic mice has previously demonstrated that donor cells can be engrafted and visualized in recipient mouse tissue for over 22 weeks after transfusion [25
, 26
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Treatment of mice with HK-BCG
As described previously [24
], the cultured bacteria of M. bovis BCG Tokyo 172 strain were washed, autoclaved, and lyophilized. This HK-BCG powder was suspended in pyrogen-free saline and dispersed by brief (10 s) sonication immediately before use. These HK-BCG preparations contained undetectable levels of endotoxin (<0.03 EU/ml), as determined by the Limulus amebocyte lysate assay (Sigma Chemical Co., St. Louis, MO) [24
]. Groups of mice (four to five/group) received 1 mg HK-BCG i.p. on Day 0. Controls received 0.2 ml saline. Spleens were harvested on Days 0, 1, 2, 3, 7, and 14.
Bone marrow transfusion
Groups of donor GFP mice (five/group) received 1 mg HK-BCG i.p. on Day 0. Controls received 0.2 ml saline. To obtain bone marrow cells for transfusion, marrow cavities of femurs were flushed with ice-cold medium RPMI 1640; the expelled cell plugs were refluxed gently with a 20-gauge needle to form a single-cell suspension. For preparation of bone marrow transfusion recipients, five WT mice, 10 weeks of age, were treated with 1 mg HK-BCG i.p. On Days 1 and 2, the WT recipients received 2 x 107 bone marrow cells isolated from BCG-treated GFP donors [18
]. In all experiments, bone marrow cells for transfusion were prepared from a pool of five donors.
Splenic MØ preparation
Spleens from each group of mice were isolated and pooled. Excised spleens were minced with scissors and digested with 50 U/ml collagenase D (C-2139, Sigma Chemical Co.) in RPMI 1640 plus 10% fetal bovine serum (FBS) at 37°C for 60 min followed by filtration through 100 µm mesh. Cells were washed with RPMI 1640 in the presence of 100 µg/ml DNase (DN-25, Sigma Chemical Co.) and resuspended in RPMI 1640 plus 10% FBS at 4 x 106 cells/ml. To enrich the MØ fraction [9
, 18
], spleen cell suspensions were layered over a discontinuous Percoll gradient (35/60%, Sigma Chemical Co.). Following centrifugation (800 g, 30 min, 22°C), cells at the interface between 35% and 60% Percoll were collected, washed, and suspended in RPMI 1640 plus 10% FBS. These cells were plated at 3 to 5 x 106 cells per 60 mm culture dish (Falcon, Oxnard, CA) and incubated at 37°C in 5% CO2 in air. After 2 h incubation, cells were washed with Ca2+- and Mg2+-free 0.15 M phosphate-buffered saline for removal of nonadherent cells. Culture dishes were placed on ice for 30 min before harvesting the adherent cells by scraping with a cell scraper (Corning, Corning, NY) and washing twice with serum-free RPMI 1640. Viability was >90% by trypan blue exclusion. Adherent spleen cells were >70% MØ, estimated by phagocytosis of immunoglobulin G (IgG)-opsonized sheep red cells and/or cytometrically following staining with anti-F4/80 [9
, 18
, 27
].
For assay of PGE2 release, plastic-adherent splenic MØ (2x106/ml) were incubated in serum-free RPMI 1640 with 10–6 M calcium ionophore A23187 (Sigma Chemical Co.) or 1 µg/ml AA (Cayman, Ann Arbor, MI) for 2 h. PGE2 levels in the culture supernatants were measured by competitive enzyme-linked immunosorbent assay (ELISA; Cayman).
For determination of IL-10 production, splenic MØ (2x106/ml) were cultured in RPMI 1640 plus 5% FBS with 100 µg/ml HK-BCG or 1 µg/ml bacterial endotoxin (LPS) for 24 h. IL-10 levels in the culture supernatants were measured by sandwich ELISA (PharMingen, San Diego, CA).
Magnetic separation of F4/80- or RB6-8C5-positive cells
Red cell-free spleen cells (108 cells) were treated with 5 µg/ml monoclonal antibody (mAb) to F4/80 recognizing splenic MØ (Accurate Chemical & Scientific Corp., Westbury, NY), followed by addition of 200 µl magnetic microbead-conjugated goat anti-rat IgG (130-048-501, Miltenyi Biotec, Auburn, CA). F4/80- positive and -negative cells were isolated according to the company’s instructions. The F4/80+ cells in positive and negative cell preparations were 87% and 0.8%, respectively (data not shown), determined cytometrically [5
]. Similarly, RB6-8C5 (Gr-1)-positive and -negative populations were also isolated using a specific antibody (Research Diagnostics, Flanders, NJ). The RB6-8C5 cells in positive and negative cell preparations were 91% and 0.9%, respectively (data not shown).
Western blotting
Splenic MØ, prepared as described above, were washed three times with cold saline. Washed cells were resuspended in lysis buffer {50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1:500 protease inhibitor cocktail (P8340, Sigma Chemical Co.), 1% Nonidet P-40, and 1% sodium deoxycholate [5
]}. Debris was eliminated by centrifugation (5 min, 1000 g). Protein concentration in the lysate was measured with a bicinchoninic acid (BCA) assay (Pierce, Rockford, IL) and bovine serum albumin as standard. Equal amounts of protein were loaded onto sodium dodecyl sulfate-polyacrylamide minigels and separated by electrophoresis (200 V for 45 min). Proteins were then transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA). The membrane was blocked with 5% nonfat dry milk and incubated with antibody [anti-COX-1, 1:1000; anti-COX-2, 1:4000; anti-cPGES, 1:1000; anti-mPGES-1, 1:1000, all from Cayman; anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH), 1:4000, from Novus Biologicals (Littleton, CO), for the detection of GAPDH as a constitutively expressed protein control] in 5% nonfat dry milk, overnight at 4°C. Following incubation with peroxidase-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch, West Glove, PA), proteins were detected by enhanced chemiluminescence (Amersham, Piscataway, NJ) following the manufacturer’s instructions [5
]. Intensity of the specific bands was quantified digitally with graphic imaging software (NIH Image 1.5).
Immunohistochemistry
Spleens were fixed with 4% paraformaldehyde overnight at 4°C, cryoprotected in 30% sucrose at 4°C for 24 h, embedded in Tris-buffered saline-freezing medium (Triangle Biomedical Sciences, Durham, NC), and stored at –70°C. Frozen spleens were cryosectioned at –20°C, thaw-mounted onto Plus-treated glass slides, and heat-set at 30°C for 3 min before air-drying overnight. For immunohistochemical analysis, 8 µm sections were stained with anti-COX-2 antibody at 1:250 (Cayman) followed by tetramethyl rhodamine-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch) at 1:100. Specific staining was detected by fluorescence microscopy (Zeiss Photomicroscope III, Carl Zeiss, Inc., New York, NY).
Cell-free PGE2 biosynthesis assay
COX and PGES activities in cell lysates were measured as conversion of exogenous AA and exogenous PGH2, respectively, to PGE2 [16
, 28
]. For these assays, F4/80+ splenic MØ in 400 µl 10 mM Tris-HCl, pH 8.0, were disrupted by sonication using a Branson sonifer (10 sx3 at 1-min intervals). After centrifugation of the sonicates at 1700 g for 10 min, 4°C, protein concentrations in the supernatants were determined by BCA assay, as described above, and adjusted to 1 mg/ml. For COX assay, 10 µl each lysate (10 µg protein) was incubated for 15 min at 37°C with 3 µM AA in 0.1 M Tris-HCl, pH 8.0, containing 1 mM reduced L-glutathione (Sigma Chemical Co.). For PGES assay, the lysate in the same buffer containing reduced L-glutathione was incubated with 0.5 µg PGH2 (Cayman) and 5 µg indomethacin for 30 s at 24°C. Reactions were terminated by addition of 100 mM FeCl2, and PGE2 in the supernatants was quantified by competitive ELISA (Cayman).
Cytometric detection
The expression of surface antigens on spleen cell preparations was determined by indirect immunofluorescence in the presence of 5% heat-inactivated newborn calf serum (Sigma Chemical Co.), pH 7.2. Rat mAb used for the analyses were membrane-activated complex-1 (Mac-1; ß2 integrin), F4/80 (red pulp MØ), ER-TR9 (marginal zone MØ), and RB6-8C5 (Gr-1, neutrophils, all from Research Diagnostics). Phycoerythrin (PE)-conjugated donkey anti-rat IgG (Jackson ImmunoResearch) was used as secondary antibody for indirect immunofluorescence.
To determine expression of cytosolic antigens, spleen cells prepared above were fixed with 4% paraformaldehyde, permeabilized with 1% saponin, and stained with rabbit antibodies specific for COX-1 or COX-2 at 2 µg/ml (Cayman) [29 ]. Presence of the primary antibody was determined by addition of PE-conjugated donkey anti-rabbit IgG at 1:1000 (Jackson ImmunoResearch). Fluorescence of 104 stained cells, unless stated, was quantitated with a FACScan flow cytometer using the CellQuest program (Becton Dickinson, Mountain View, CA). All cells, as defined by forward- and sideward-scatter pattern, were gated; only debris was excluded from analysis. Cells stained with the secondary antibody alone were used as negative controls in all experiments.
Statistics
Data for PGE2 release were analyzed by one-way ANOVA. For cell culture studies, tissues isolated from at least four mice were pooled unless indicated; these cells were cultured in at least triplicate in each group. P < 0.05 is considered statistically significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (21K): [in a new window] |
Figure 1. HK-BCG induces splenic PGE2-MØ formation in IL-10–/– and C57Bl/6 (WT) mice. Groups of IL-10–/–and C57Bl/6 (WT) mice (four/group) received 1 mg HK-BCG i.p. on Day 0. At indicated intervals, spleens were harvested, and plastic-adherent splenic MØ were isolated and pooled in each group. To determine PGE2 biosynthesis, MØ suspended in serum-free RPMI 1640 were incubated with saline (open bars), 1 µg/ml AA (shaded bars), or 1 µM A23187 (solid bars) for 2 h. PGE2 in the supernatants was measured by ELISA. Mean ± SD, n = 3. *, P < 0.05, compared with Day 0 control group. Results shown are representative of results from three separate experiments.
|
 View larger version (47K): [in a new window] |
Figure 2. COX-1, COX-2, cPGES, and mPGES-1 levels in HK-BCG-treated mice. IL-10–/– and WT mice received 1 mg HK-BCG in saline, i.p., on Day 0, and Days 0 (untreated), 1, 2, 3, 7, and 14 spleens were harvested. Plastic-adherent splenic MØ were isolated from each animal and analyzed by Western blotting. GAPDH was included as a constitutively expressed control.
|
 View larger version (28K): [in a new window] |
Figure 3. COX-2 levels in F4/80+ cells. F4/80+ and F4/80– cells were isolated from Days 1 and 14 whole spleen nucleated cells obtained from WT and IL-10–/– mice as indicated in Materials and Methods. (A) COX-2 and GAPDH expressions in each fraction were determined by Western blotting. (B) Intensity of COX-2 bands indicated in A was quantified digitally using graphic imaging software (NIH Image 1.5) and is shown as ratio to GAPDH intensity. Mean ± SD, n = 3. *, P < 0.01, compared with WT F4/80+ group. Results are representative of two separate experiments.
|
3 µM) from WT F4/80+ or IL-10–/– F4/80+cells compared with untreated controls. In contrast, on Day 14 after treatment, there was a significant increase in PGE2 release by WT and IL-10–/– MØ (Fig. 4A)
. The PGE2 produced by IL-10–/– cells on Day 14 was significantly greater than that for WT MØ and correlated with the increased COX-2 levels. Exogenous AA may serve directly as a COX substrate, independent of PLA2 activity. However, exogenous AA is also reported to increase the level of intracellular Ca2+ [30
], resulting in activation of PLA2. Therefore, to assess further the COX and PGES enzymatic activities in splenic MØ, cellular lysates were mixed with exogenous AA or PGH2, as substrates for COX or PGES, respectively. Day 1 cellular lysates did not convert exogenous AA to PGE2 in either strain of mice (Fig. 4B)
. PGES activity was present in untreated cells and was not altered significantly by treatment with HK-BCG (Fig. 4C)
on Day 1 or 14, strongly suggesting that COX activity is rate-limiting for PGE2 biosynthesis. Thus, the COX catalytic activity in Day 1 MØ is disproportionate to the COX protein level, which would account for the observation that PGE2 production is not increased on Day 1.
 View larger version (21K): [in a new window] |
Figure 4. PGE2 biosynthesis, cell-free COX, and PGES activities in F4/80+ cells from HK-BCG-treated mice. F4/80+ cells were isolated from Days 0, 1, and 14 following BCG treatment and pooled in each group, as indicated in Materials and Methods. (A) PGE2 determined following treatment of cells with AA as described for Figure 1
. Mean ± SD, n = 3. (B) COX and (C) PGES activities in cell lysates were measured as described in Materials and Methods. PGE2 was measured by ELISA. Mean ± SD, n = 3. *, P < 0.05; **, P < 0.01; #, P < 0.001, for IL-10–/– compared with WT. In the absence of exogenous substrates (AA or PGH2), endogenous PGE2 levels were  1 ng/10 µg protein (data not shown).
|
 View larger version (13K): [in a new window] |
Figure 5. Production of IL-10 by WT splenic MØ. Plastic-adherent splenic MØ were isolated Days 0, 1, and 14 following BCG treatment of WT and IL-10–/– mice as indicated in Materials and Methods. MØ (2x106/ml) were stimulated with 100 µg/ml HK-BCG, 1 µg/ml LPS, or medium at 37°C for 24 h. IL-10 in culture supernatants was measured by ELISA. Mean ± SD, n = 3; *, P < 0.001, compared with cells from the same day treated with saline. The data shown are representative of two independent experiments.
|
Bone marrow cells, spleen cells, and peritoneal cells were prepared from HK-BCG-treated chimeras, and GFP expression was determined cytometrically. Figure 6 shows that small but significant numbers of GFP+ cells were detected in the tissues isolated from the chimera (Table 1 ). The chimera formed in animals treated with HK-BCG resulted in higher engraftment of donor cells than in the untreated chimera (Fig. 6B and 6D) . Our results suggest that treatment of mice with HK-BCG was essential to induce tissue localization of donor cells. This was most evident in the peritoneal cells (Fig. 6) . Selected splenic MØ populations, characterized by mAb against Mac-1, F4/80, and ER-TR9 and derived from GFP+ cells, were identified in the chimera (Table 1) . As numbers of donor bone marrow cells transfused are only a fraction of bone marrow cells in the recipients, GFP+ cell numbers localized in the chimeric tissues appear to be relatively small, even with HK-BCG treatment. For the HK-BCG-treated GFP chimeric mice, GFP+ donor cells localized in the spleen were 1.27% of the total cells. Two-color cytometric analyses showed that 0.29% of the total spleen cells were COX-2+ and GFP+, whereas 10.67% were COX-2+ and GFP–, suggesting that the COX-2+ cells originated mainly from recipient cells (Table 1 , Fig. 7 ). Figure 8 illustrates immunohistologically the colocalization of COX-2 and GFP in spleen cells of BCG-treated mice following bone marrow transfusion.
 View larger version (15K): [in a new window] |
Figure 6. Chimeras for bone marrow, spleen, and peritoneal cells in C57Bl/6 recipients, which received bone marrow cells from GFP-transgenic donors. WT recipients and GFP-transgenic donors were immunized with HK-BCG. As controls, recipients and donors were immunized with saline. Donor bone marrow cells (2x107) were transfused into the recipients on Days 1 and 2 after immunization. On Day 14, recipients were killed, and bone marrow, spleen, and peritoneal lavage cells were isolated. GFP levels were determined cytometrically. As controls, cells were prepared from WT recipients and GFP-transgenic donors, 14 days after HK-BCG immunization. (A) HK-BCG-treated WT recipient and GFP donor controls; (B) HK-BCG-treated WT recipient control and GFP-chimera; (C) untreated WT recipient and GFP donor controls; and (D) untreated WT recipient control and GFP-chimera.
|
|
View this table: [in a new window] |
Table 1. Percent of Spleen Cells Expressing Selected Cellular Antigens in the HK-BCG-Treated GFP-Chimera
|
 View larger version (17K): [in a new window] |
Figure 7. Cytometric detection of spleen cells expressing GFP and COX-2 in the HK-BCG-treated GFP-chimera. GFP-chimeric mice following treatment with HK-BCG were prepared as described in Materials and Methods. Spleen cells were isolated from (A) GFP-chimera, (B) WT recipients, and (C) GFP-donor controls. Intracellular COX-1 and COX-2 were stained with anti-COX-1 and anti-COX-2, followed by PE-conjugated donkey anti-rabbit IgG. Normal rabbit IgG served as a negative control for primary antibody binding. Fluorescence of 105-stained cells was quantitated cytometrically. PE-positive cells and GFP-positive cells were counted by two-color analysis.
|
 View larger version (30K): [in a new window] |
Figure 8. Immunofluorescent detection of spleen cells expressing GFP and COX-2 in the HK-BCG-treated bone marrow chimera. HK-BCG-treated GFP-chimera was established as described in Materials and Methods. On Day 14 after HK-BCG treatment, spleens of the chimera were harvested, fixed, frozen, and sectioned as described in Materials and Methods. For detection of COX-2+ cells, sections were stained with anti-COX-2 primary antibody followed by tetramethyl rhodamine (red)-conjugated secondary antibody. In the spleen cell sections shown at 100x original magnification, GFP (green) was colocalized with COX-2 (red). Without HK-BCG treatment in the donors and recipients, there were no COX-2+ cells or GFP+ cells in the spleen (data not shown).
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
An explanation for the dissociation between expression of COX-2 protein by Day 1 splenic MØ and the COX-2 activity indicated by the unchanged PGE2 biosynthesis is not apparent. Analogous to the in vivo treatment with HK-BCG, our previous in vitro studies with WT splenic MØ treated with LPS and IFN- resulted in increased COX-2, apparently independent of IL-10, but the release of PGE2 was only increased minimally [21
]. Certain tumor cells also express high levels of COX-2 but little PGE2, as a result of catalytically inactive COX-2 [31
]. As PGES activity is relatively constant in our cells, our results (Fig. 4)
also suggest that COX-2 induced on Day 1 following HK-BCG treatment lacks catalytic activity, possibly associated with specific post-translational regulation. It is known that COX activation requires a small amount of peroxide, which may be supplied by peroxynitrite formed from superoxide anion and nitric oxide [32
]. It is possible that the necessary peroxide is not present in the Day 1 cells. Alternatively, PGE2 may not be synthesized because of inadequate coupling of COX-2 with a PLA2 providing substrate or a terminal PGES. However, there do not appear to be previous reports clearly demonstrating this type of regulation of COX activity [33
]. Regardless of the mechanism, the lack of PGE2 biosynthesis by Day 1 COX-2+splenic MØ would imply that unlike COX-2+splenic MØ on Days 7–14, Day 1 COX-2+ MØ do not regulate immune responses in a PGE2-dependent manner.
Regulation of COX-2 expression by IL-10 appears to occur at the level of COX-2 mRNA. Berg et al. [20
] found that LPS stimulation of spleen cells from IL-10–/– mice resulted in a level of COX-2 mRNA, which was 5.6-fold greater than in WT cells and correlated with increased COX-2 protein and PGE2 production. In the IL-10–/– cells, the effect of LPS on COX-2 induction was reversed in the presence of neutralizing, anticytokine antibodies to IFN-, tumor necrosis factor 
, or IL-12, which suggests that the increase in COX-2 resulted from increased levels of these proinflammatory cytokines [20
]. Thus, IL-10 may act indirectly by regulating the levels of these cytokines. In U937 cells, COX-2 mRNA was stabilized following treatment with LPS or IL-1ß [34
]. In addition, previous studies [35
36
37
] suggest that LPS-induced COX-2 expression and PGE2 synthesis are associated with p38 mitogen-activated protein kinase (MAPK) activation in human monocytes and neutrophils. The activity of p38 MAPK was necessary for stabilization of COX-2 mRNA [35
36
37
]. IL-10 appears to down-regulate p38 MAPK activation [38
39
40
]. p38 MAPK is an upstream kinase regulating nuclear factor-
B activation in neutrophils [41
], suggesting that p38 MAPK might play a role in transcriptional and post-transcriptional regulation of the COX-2 gene. Further studies will be required to dissect the mechanisms of IL-10-dependent regulation of the expression of COX-2.
Different populations or compartments of MØ show considerable diversity in PGE2 biosynthesis. Using mice depleted of bone marrow cells and circulating monocytes by the bone-seeking isotope 89Sr, we demonstrated previously that the development of splenic PGE2-MØ is dependent on radio-sensitive bone marrow [9 , 18 ] but that PGE2-MØ development in the peritoneal cavity is independent of the bone marrow. Our results (Fig. 6) clearly show that transfusion of WT recipients with GFP+ bone marrow cells establishes a GFP-chimera in bone marrow, spleen, and peritoneal cells following HK-BCG treatment. Thus, splenic PGE2-MØ may be derived directly from the bone marrow cells transfused on Days 1 and 2 after HK-BCG treatment, and bone marrow does not solely supply an appropriate cytokine milieu. Taken together, it is likely that at the early stages (1–2 days) after HK-BCG treatment, precursors of PGE2-MØ in the bone marrow are induced to migrate and localize in the spleen where mature forms of PGE2-MØ are established [9 , 18 ]. Thus, COX-2+ MØ, present on Days 1 and 14, may be derived from distinct populations of cells, which may account, at least in part, for the deficient PGE2 synthesis observed on Day 1. Days 2 and 3, when COX-2 is not expressed, appear to be a transition from a transient expression of COX-2 by local MØ to the Day 7 COX-2 expression by MØ newly derived from bone marrow.
Increases in immature and mature MØ and neutrophil numbers within 5–21 days are observed in animals treated with BCG or P. acnes [9 , 42 ]. Our unpublished studies have also shown that depletion of RB6-8C5+ cells by a specific antibody on Day 14 in HK-BCG mice results in unchanged F4/80+ red pulp MØ numbers as well as PGE2 release (data not shown). In addition, on Day 14, COX-2 was not expressed differentially on cells positive or negative for RB6-8C5 or Mac-1 (data not shown). However, ER-TR9+ marginal zone MØ, distinct from F4/80+ red pulp MØ, potentially could be PGE2-MØ. The possible expression of COX-2 by DC remains to be examined. Thus, our results do not rule out the possibility that HK-BCG induces cell types other than F4/80+ MØ, which contribute to the release of PGE2.
In conclusion, our previous finding [18 ] that development of splenic PGE2-MØ is dependent on radio-sensitive bone marrow and that these cells are probably derived directly from bone marrow precursors is confirmed further. Splenic PGE2-MØ populations persist for long periods, therefore, prolonging the effect of PGE2 on immune regulation including the Th1-to-Th2 shift of immune responses against mycobacterial antigens [5 ]. Although resident splenic MØ can be induced to express COX-2, these cells do not appear to produce increased levels of PGE2. We find further that endogenous IL-10 contributes to the regulation of COX-2 expression and the resultant PGE2 production in cells obtained 7–14 days following treatment with BCG. Thus, our study indicates that the generally accepted concept that resident COX-2– MØ are converted to COX-2+ MØ under inflammatory conditions with release of relatively large amounts of PGE2 may need further investigation, particularly with respect to splenic MØ.
Received December 15, 2005; revised April 7, 2006; accepted May 2, 2006.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-induced macrophage NO synthesis by PGE2 J. Leukoc. Biol. 61,712-720[Abstract] and interleukin-10 production by prostaglandins I(2) and E(2): studies with prostaglandin receptor-deficient mice and prostaglandin E-receptor subtype-selective synthetic agonists Biochem. Pharmacol. 61,1153-1160[CrossRef][Medline] J. Immunol. 163,963-969 mitogen-activated protein kinase in lipopolysaccharide-stimulated neutrophils J. Clin. Invest. 103,851-858[Medline]This article has been cited by other articles:
 |
|
 |
|
  K. Bansal, Y. Narayana, S. A. Patil, and K. N. Balaji M. bovis BCG induced expression of COX-2 involves nitric oxide-dependent and -independent signaling pathways J. Leukoc. Biol., May 1, 2009; 85(5): 804 - 816. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Nishiyama, T. Shinohara, T. Pantuso, S. Tsuji, M. Yamashita, S. Shinohara, Q. N. Myrvik, R. A. Henriksen, and Y. Shibata Depletion of cellular cholesterol enhances macrophage MAPK activation by chitin microparticles but not by heat-killed Mycobacterium bovis BCG Am J Physiol Cell Physiol, August 1, 2008; 295(2): C341 - C349. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Yamashita, T. Shinohara, S. Tsuji, Q. N. Myrvik, A. Nishiyama, R. A. Henriksen, and Y. Shibata Catalytically Inactive Cyclooxygenase 2 and Absence of Prostaglandin E2 Biosynthesis in Murine Peritoneal Macrophages following In Vivo Phagocytosis of Heat-Killed Mycobacterium bovis Bacillus Calmette-Guerin J. Immunol., November 15, 2007; 179(10): 7072 - 7078. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Yamashita, S. Tsuji, A. Nishiyama, Q. N. Myrvik, R. A. Henriksen, and Y. Shibata Differential subcellular localization of COX-2 in macrophages phagocytosing heat-killed Mycobacterium bovis BCG Am J Physiol Cell Physiol, July 1, 2007; 293(1): C184 - C190. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
