Published online before print July 25, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Ludwig Institute for Cancer Research, Melbourne Tumor Biology Branch, Melbourne, Victoria, Australia;
Walther Oncology Center, Indiana University School of Medicine, Research Institute No. 2, Indianapolis, Indiana, USA;
Commonwealth Serum Laboratories, Parkville, Victoria, Australia; and
Amgen Inc., Thousand Oaks, California, USA
1 Correspondence: Ludwig Institute for Cancer Research, P.O. Box 2008 RMH, Melbourne, Victoria 3050, Australia. E-mail: francesca.walker{at}ludwig.edu.au
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: cytokines • granulopoiesis • knockout mice
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The pivotal role of G-CSF and GM-CSF in "in vivo" granulopoiesis has been demonstrated through pharmacological administration [4 ] or by genetic ablation of the corresponding genes [5 , 8 9 10 11 ]. However, recent findings using G-CSF/GM-CSF double-deficient mice (G–/–, GM–/–) have suggested the existence of other regulators of neutrophil production. In previous studies, we have demonstrated that although G-CSF-deficient mice are neutropenic, they develop a profound neutrophilia when challenged with Candida albicans [12 ]. Similarly, an increase in neutrophil numbers in peripheral blood is observed in GM-CSF/G-CSF double-knockout (KO) mice and in G-CSF/IL-6 double-KO mice challenged with C. albicans for 7 days [12 , 13 ]. As these mice cannot produce any biologically active G-CSF or GM-CSF [5 , 11 ], these observations suggest that the proliferation and differentiation of bone marrow cells at the GM-progenitor level may be stimulated by another cytokine pathway. Indeed, we have previously demonstrated enhanced granulopoiesis "in vitro" when stromal cells from G-CSF and GM-CSF double-deficient mice are exposed to C. albicans [13 ].
To extend these findings and to characterize the putative neutrophil-promoting activity (NPA), we have developed a robust liquid culture assay system for the detection of NPA in media conditioned by primary and immortalized embryonic fibroblasts from G–/–, GM–/– mice. We show that the level of NPA in culture medium conditioned by embryonic fibroblasts is strongly enhanced after stimulation of the cells with pathogen(s). Using embryonic fibroblasts from G-CSF/GM-CSF/ M-CSF triple-KO mice (G–/–, GM–/–, M–/–) or blocking anti-M-CSF antibodies, we have uncovered an obligate role for M-CSF in the production of NPA.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and IL-12 were from R&D Systems (Minneapolis, MN, USA). LPS from Escherichia coli 0127:B8 was purchased from Sigma Chemical Co. (St. Louis, MO, USA). Vascular endothelial growth factor [VEGF; affinity-purified from the conditioned medium (CM) of VEGF-expressing Chinese hamster ovary (CHO) cells] was a kind gift from Maria Macheda (Angiogenesis Laboratory, Ludwig Institute, Melbourne, Australia). Polyclonal anti-SCF receptor antibody was a 10x concentrated supernatant from the Ack hybridoma [14 ]. Goat anti-mouse-neutralizing antibodies to M-CSF (M9082) and to SCF antibody (S-6045) were purchased from Sigma Chemical Co. Goat anti-mouse Cytokeratine-19 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-c-fms AFS98, anti-c-Kit (CD117, PE-coupled), anti-fetal liver tyrosine kinase 3 (Flt-3; CD135), anti-membrane-activated complex 1 (Mac-1; CD 11b, FITC-coupled), and anti-granulocyte-differentiation antigen-1 (Gr-1) Ly6 G+C, PE-coupled) were purchased from BD Biosciences (Franklin Lakes, NJ, USA). Heat-inactivated (HI) C. albicans was prepared as described previously [12 ].
Animals
G-CSF–/–/GM-CSF–/– mice [10
] and G-CSF–/–/GM-CSF–/–/M-CSF–/– mice (M. L. Hibbs, C. Quilici, N. Konfouri, T. F. Seymour, J. E. Armes, A. W. Burgess, A. R. Dunn, manuscript submitted) were used for these studies. GM–/–G–/–M+/– mice were generated by intercrossing osteopetrotic (op/op) mice [15
] with GM–/–G–/– mice. These were, in turn, intercrossed to obtain G, GM, M–/– embryos. All animal experiments were conducted according to the guidelines of the National Health and Medical Research Council of Australia (Canberra). All animal procedures were approved by the Ludwig Institute (Melbourne, Australia) and Department of Surgery Animal Ethics Committee under License 27/02.
Primary murine embryonic fibroblast (MEF) culture
G–/–GM–/– or G–/–GM–/–M+/– pregnant female mice were killed on the 13th–15th day of pregnancy by cervical dislocation. Embryos were washed with PBS. Soft tissues (i.e., liver, heart, and other viscera) were removed, and embryonic carcasses were digested in trypsin at 37°C for 30 min followed by overnight incubation at 4°C. Embryonic fibroblasts were cultured in DMEM containing 15% HI FCS. When confluent, cells were subcultured at a dilution of one in three. Cells from individual embryos were also collected at this time for genotyping.
Immortalization of embryonic fibroblasts
Embryonic fibroblasts were immortalized by transfection with SV40 T vector pT22 [16
], from Professor Phil Gallimore (University of Birmingham, UK). Primary embryo fibroblasts were passaged 24 h before electroporation to ensure logarithmic growth. Cotransfections were carried out using a 20:1-µg ratio of pSV40-T plasmid DNA and pTK-hyg plasmid DNA (Clontech, Palo Alto, CA, USA). Cells were transfected by electroporation in a 0.4-cm cuvette, pulsed at 260V/960µF using a gene pulser (Bio-Rad Laboratories, Hercules, CA, USA). Forty-eight hours after electroporation, cells were passaged and selected in 100 µg/ml hygromycin B (Roche, Dee Why, New South Wales, Australia). Transfected cells were subcloned further in soft agar. Stable expression of SV40-T was confirmed by PCR.
Detection of SV40 large T antigen
Specific primers for SV40 large T antigen [17
] were sense: 5' CCT GGC TGT CTT CAT CAT C 3'; antisense: 5' CAT GAA TGA GTA CAG TGT GCC 3'. PCR analysis was performed on DNA from hygromycin-selected clones and yielded a product of 340 bp.
PCR diagnosis of M-CSF genotype
Mice and embryos were genotyped at the M-CSF locus by PCR analysis. The primers to identify the op allele were designed to amplify the segment containing the op mutation on Chromosome 3 [8
]. The two primers used were sense: 5' GAT CCT GTT TGC TAC CTA AAG AAG GCC ATT TT 3'; antisense: 5' GTA AGG TTC AGG CTC GGT GAG CAT AT 3'. The PCR products were digested with BSTX I at 45°C and run on a 4% agarose gel (Reliant 4% Tris-boric acid-EDTA buffer gel, MC54929). Under these conditions, the wild-type (M+/+) produces two bands of 30 bp and 230 bp; homozygous (op/op, M–/–) only produces a 263-bp band, and heterozygous (op/+, M+/–) produces three bands of 263 bp, 233 bp, and 30 bp.
Production of CM
Primary culture (P1) MEFs were passaged at 1:3 dilution. After overnight incubation, passage 2 (P2) embryonic fibroblasts were stimulated with LPS (0.05 µg/ml) or with 1 x 108 HI C. albicans (ATCC 18804) per 175 cm2 flask in phenol red-free DMEM (Gibco, Invitrogen, Carlsbad, CA, USA), supplemented with 1% human albumin solution (CSL, Australia) and 1 mM sodium pyruvate (Gibco, Invitrogen). CM were harvested after 48 h.
SV40 T-transfected cell lines, seeded at 1 x 107 cells per 175 cm2 flask the day before stimulation, were stimulated with LPS or HI C. albicans as described in the preceding paragraph, and CM were harvested after 48 h.
Bone marrow assay
GM–/–G–/– mice were injected i.p. with 4 mg cyclophosphamide (Sigma Chemical Co.) in 0.2 ml PBS 3 days before assay to enrich for myeloid progenitors. Femora from at least two mice were used for each experiment. The bone marrows were flushed into 5 ml low-density (1.060 g/ml) metrizamide (ICN, Costa Mesa, CA, USA; Cat. #195324 [18
]) or Histopaque (Sigma Chemical Co.) and dispersed to yield a single-cell suspension, which was layered over high-density (1.080 g/ml) metrizamide or Histopaque solution and centrifuged at 1500 g for 15 min at 4°C. The band of low-density cells at the interface was collected and washed with culture medium. Cells were cultured in eight-well LabTek chamber slides (Nunc, Rochester, NY, USA) at 5 x 103 cells/well in a total volume of 200 µl 1:1 RPMI and IMEM, supplemented with 10% HI FCS and nonessential amino acids (all Gibco products from Invitrogen). After 4 days in culture, the slides were stained with May-Grumwald Giemsa for 4 min followed by 3% Giemsa's stain (both from BDH, Merck, Darmstadt, DL, Germany) for 7 min. Granulocyte clusters, containing between five and 50 cells, were identified by morphology or by specific staining with naphthol AS-D chloroactetate esterase. NPA activity is expressed in units/ml. One unit is defined as the activity required to generate one granulocyte cluster from 5 x 103 bone marrow cells after 4 days in this liquid culture assay.
Naphthol AS-D chloroacetate esterase staining
Slides were air-dried, fixed in formalin vapor for 4 min, rinsed in water, and stained using the Naphthol AS-D chloroacetate (specific esterase) kit (Sigma Chemical Co.), according to the manufacturer's instructions. Cells were counterstained with hematoxylin.
Microscopy
Slides were examined using a Nikon Microphot-FX microscope. Images were acquired using a 20x plan or 100x plan (oil) lens and a diagnostic imaging spot camera.
FACS analysis
Bone marrow cells from cyclophosphamide-treated G–/–GM–/– mice were harvested and enriched for myeloid progenitors as described above. The cells were divided into three samples: One was processed directly for antibody staining and FACS analysis, and the other two were transferred to flasks and cultured with or without CM from C. albicans-stimulated G–/–GM–/– embryonic fibroblasts as a source of NPA. Cultured cells were harvested after 4 days and processed for antibody staining and FACS analysis. Single-cell suspensions (0.5–2x106 cells per tube) were incubated with a combination of FITC-labeled anti-Mac-1 (M1/70) and PE-labeled anti-Gr-1 (Rb6-8C5A) or FITC-labeled anti-Flt-3 and PE-labeled anti-Kit antibodies in FACS buffer (5% FCS, 5% EDTA in mouse toxicity PBS) for 20–30 min. Washed cells were analyzed on a FACScan (Becton Dickinson, North Ride, New South Wales, Australia). Nonspecific binding of antibodies to FcRs (Fc
III/II) was reduced by preincubating the cells with Fc-block (BD Biosciences PharMingen, San Diego, CA, USA) for 10 min on ice. The data were plotted and analyzed using the statistical software in CellQuest
Use of neutralizing antibodies during NPA production and in bone marrow assays
To monitor the requirement for M-CSF during NPA production, SV40 T-transformed cells (SV#23) were stimulated with LPS (0.05 µg/ml), alone or combined with anti-M-CSF antibody (M9082) or anti-c-fms antibody (AFS98) at different concentrations. Parallel cultures of SV#43 cells were incubated with LPS and control antibodies to mouse SCF (S-6045) or to cytokeratine 19 (SC-33119): These antibodies are polyclonal antibodies produced in goats and are matched to the M-CSF antibody M9082. After 48 h culture, CMs were harvested and tested in bone marrow assays.
To block growth factor receptors on the target bone marrow cells, cells were incubated with anti-c-fms antibody (AFS98) or with anti-SCF receptor antibody (ACK-2; 10x concentrated from hybridoma supernatants) in a chamber slide at 37°C for 30 min before adding the growth factors or CM. To neutralize the action of growth factors, blocking antibodies (anti-M-CSF antibody M9182) were incubated with growth factors or CM at 37°C for 30 min in a chamber slide before adding the bone marrow cells. Culture conditions, staining of the slides, and scoring of granulocyte clusters were performed as described previously [13 ].
Supplementation of MEF cultures with M-CSF
Primary embryonic fibroblasts from G–/–GM–/–M–/– embryos (obtained by intercrossing G–/–GM–/–M+/– mice) were stimulated with LPS (0.05 µg/ml), alone or combined with M-CSF (5 ng/ml). The CMs were harvested 48 h later and tested in the bone marrow assay. For some experiments, suspensions of embryonic fibroblasts were divided into two parts immediately after isolation from embryo bodies. One part was cultured in normal medium, the other, in medium containing M-CSF (5 ng/ml). One day before stimulation, cells from each of the cultures were trypsinized, washed, and seeded at 1 x 105 cells/well in 24-well plates and then, stimulated with LPS (5 µg/ml), alone or combined with M-CSF (5 ng/ml). The CMs were harvested after 48 h culture and tested in a bone marrow assay.
For SV40 T-immortalized G–/–GM–/–M–/– (T-SV cell line), M-CSF (5 ng/ml) was kept in the culture medium throughout the isolation of the cells from embryos, transfection, and selection in hygromycin. To obtain the CM, these cells were stimulated with LPS alone or combined with M-CSF.
Cytokine array analysis
Embryonic fibroblasts from G–/–GM–/–M+/– or G–/–GM–/–M–/– embryos were incubated in medium, with or without LPS (0.05 µg/ml), as described in Production of CM above. The CM were collected after 48 h. RayBio mouse cytokine antibody array membranes (Array II, RayBiotech, Norcross, GA, USA) were incubated with the individual CM overnight and then processed according to the manufacturer's instruction manual. Reactivity with individual antibodies was visualized with ECL reagent (Amersham, Piscataway, NJ, USA). Multiple exposures of each film were obtained to ensure linearity of detection.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We have now developed a liquid culture assay to detect NPA in CM. Bone marrow cells from cyclophosphamide-treated G–/–GM–/– mice, enriched for myeloid progenitors by metrizamide density gradient [18
], were cultured on chamber slides with purified cytokines or CM from double-KO (G–/– GM–/–) MEFs. After 4 days, the number and type of myeloid cell clusters were assessed by morphology (Fig. 1a
1b
1d
1f
and 1h
) and by staining for specific esterase (Fig. 1c
1e
1g
and 1i)
. -Naphthol AS-D choloroacetate esterase staining shows localized, blue granular staining in the cytoplasm of nearly all neutrophilic granulocytes, and monocytes, macrophages, and lymphocytes do not react [19
]. The specificity of this stain is clearly seen in Figure 1e
, where macrophages and monocytes show no reactivity, and cells morphologically identifiable as granulocytes show strong positivity for this stain.
 View larger version (47K): [in a new window] |
Figure 1. Generation and identification of granulocytic clusters in liquid culture. Bone marrow cells from double-KO (G-CSF–/–, GM-CSF–/–) mice, enriched for myeloid progenitors, were cultured in eight-well chamber slides in control medium or in medium supplemented with no supplement (a); M-CSF, 10 ng/ml (b and c); GM-CSF, 10 ng/ml (d and e); SCF, 100 ng/ml (f and g); NPA, 10% v/v (h and i). (a, b, d, f, h) Images were obtained from slides stained with May-Grumwald Giemsa at 40x original magnification. (c, e, g, i) Images were taken from slides stained with -naphthol-AS-D chloroacetate at 100x original magnification.
|
We also performed FACS analysis to monitor the generation of specific blood cell types in the liquid cultures. Bone marrow cells from G–/–GM–/– mice collected immediately after metrizamide separation (before culture) or collected after 4 days in liquid culture, with or without CM, were stained with lineage-specific antibodies: Expression of Kit and Flt-3 was used to identify stem cells/early progenitor cells and Mac-1 and Gr-1 to identify cells in the myeloid/monocyte lineage [20 ]. Antibody reactivity was assessed by FACS analysis (Fig. 2 ). After culture of the cells in NPA, there was a substantial increase in the proportion of cells positive for Flt-3 and doubly positive for Flt-3 and Kit (cf., Fig. 2A and 2B ), suggesting an expansion of the early hematopoietic progenitors with myeloid proliferation potential [21 , 22 ]. It is most striking that NPA strongly enhanced the proportion of Gr-1/Mac-1 double-positive cells (Fig. 2B) , supporting the notion that under the influence of NPA, granulocytes and monocytes are the main cell types generated in liquid culture. In contrast, culture of the cells in control medium resulted in cell death (data not shown), with few surviving cells of myeloid origin (Fig. 2C) .
 View larger version (47K): [in a new window] |
Figure 2. Myeloid cells are generated during in vitro culture with CM from G-CSF/GM-CSF-deficient fibroblasts. Progenitor-enriched bone marrow cells were tested before in vitro culture (A), after 4-day culture in medium conditioned by double-KO MEFs exposed to C. albicans (B), or after 4 days culture in control medium (C). Cells were analyzed by FACS after staining with FITC (abscissa)- or PE (ordinate)-labeled antibodies. Quadrants for each population were set using FITC- or PE-labeled irrelevant antibodies, and the proportion of cells in each quadrant was determined using CellQuest.
|
 View larger version (20K): [in a new window] |
Figure 3. Exposure to C. albicans significantly enhances NPA production by G–/–GM–/– double-KO mice. G–/–,GM–/– primary embryonic fibroblasts (A) or immortalized embryonic fibroblasts (B) were stimulated with (solid bars) or without (open bars) HI C. albicans for 48 h, and the resultant CM were tested in a bone marrow assay. The number of granulocyte clusters in each chamber was determined by light microscopy on May-Grumwald Giemsa-stained slides and converted to NPA U/ml as defined in Materials and Methods. Results are mean and SD of duplicate samples (*, P<0.05; **, P<0.005; ***, P<0.001).
|
LPS enhances NPA production significantly by embryonic fibroblasts
LPS has various biological effects on mammalian hosts and their cells, one of which is to stimulate the secretion of many biologically active cytokines and mediators [24
]. We compared LPS to HI C. albicans for the induction of NPA secretion in SV40-immortalized cell lines. LPS was a much stronger stimulus for NPA production than C. albicans (Fig. 4A
), and granulocyte clusters stimulated by LPS CM were more numerous and larger (Fig. 4C)
. The identity of the cell types generated with LPS CM was confirmed by FACS analysis and -naphthol AS-D chloroacetate esterase staining.
 View larger version (35K): [in a new window] |
Figure 4. LPS enhances NPA production. (A) SV40 T-transfected cells (Clones #14 and #23, represented by open and solid bars, respectively) were exposed to medium alone, LPS (0.05 µg/ml), C. albicans, or a combination of C. albicans and LPS for 48 h. CM were harvested and tested for NPA activity on bone marrow cells. (B) CM prepared without LPS (–) or with LPS (LPS) was tested in bone marrow cultures without (open bars) or with (solid bars) further addition of 0.05 µg/ml LPS directly in the culture. (C) Cell types generated by the CM. Panels show representative fields from slides stained by May-Grumwald Giemsa.
|
M-CSF is required for NPA production
The main target cells activated by LPS are monocytes and macrophages [26
], and macrophages are a major source of many cytokines involved in the immune responses, hematopoiesis, inflammation, and other homeostatic processes [27
]. LPS has been reported to enhance the production of M-CSF by human marrow stromal cells and monocytes [27
, 28
] and of GM-CSF by monocytes [29
]. Although M-CSF action is restricted mainly to the macrophage lineage [30
], M-CSF induces the synthesis of several growth factors, including IL-1, G-CSF, IFN, and TNF [26
]. It was of interest therefore to test the role of M-CSF in the production of NPA by LPS-stimulated embryonic fibroblasts. We obtained MEFs of different genotypes from intercrosses between G–/–GM–/–M+/– mice (M. L. Hibbs, C. Quilici, N. Konfouri, T. F. Seymour, J. E. Armes, A. W. Burgess, A. R. Dunn, manuscript submitted) and assessed their ability to produce NPA when exposed to LPS (Fig. 5
). NPA levels in the CM correlated with M-CSF gene dosage: On a null background for G-CSF and GM-CSF, M-CSF+/– MEFs produced an intermediate level of NPA compared with M-CSF+/+, and NPA activity from M-CSF–/– MEFs was undetectable (Fig. 5A)
.
 View larger version (18K): [in a new window] |
Figure 5. Production of NPA is impaired in cells lacking the M-CSF gene and cannot be restored by exogenous M-CSF. Embryos were obtained from intercrosses of mice homozygote null for G-CSF and GM-CSF and heterozygote null for M-CSF. Embryos were genotyped, and primary embryonic fibroblasts were established from individual embryos. (A) MEF cultures prepared from embryos of the indicated genotype were stimulated with LPS (0.05 µg/ml) for 48 h, and the resulting CM were tested for NPA activity. All embryos were from the same litter. The data represent means and SD of duplicate samples in three separate experiments. Data from embryos of equal genotype are pooled. (B) MEF cultures established from individual embryos in the same litter were stimulated for 48 h with LPS alone (0.05 µg/ml, open bars) or combined with 5 ng/ml murine M-CSF (solid bars) and the resulting CM tested for NPA activity. The data represent means and SD of duplicate samples in three separate experiments. (C) Primary embryonic fibroblasts from G–/–GM–/–M–/– mice were derived and immortalized in medium containing M-CSF. The resulting cell line (T-SV#22) was maintained in M-CSF. T-SV#22 (G–/–, GM–/–, M–/–) and SV#23 (G–/–, GM–/–) cells were exposed for 48 h to control medium, LPS (0.05 µg/ml), or C. albicans in the presence (solid bars) or absence (open bars) of exogenous M-CSF, and the resulting CM were tested for NPA activity. Data represent means and SD of duplicate samples in three separate experiments.
|
One explanation for low NPA production by G–/–GM–/–M–/– MEFs, even when M-CSF was included in the culture, could be selective loss of a M-CSF-dependent cell type responsible for NPA production. To rescue putative M-CSF-dependent cells, we therefore added M-CSF to the MEF culture from the time of embryo isolation. As soon as embryos were dissected and trypsinized, each embryo's fibroblast was divided into two parts and cultured in medium containing 5 ng/ml M-CSF or in normal medium. When the cells were cultured for the production of CM, each set of cells was divided further into two groups: one stimulated with LPS alone and the other stimulated with LPS plus M-CSF. The resultant CM was then compared for NPA activity in the bone marrow progenitor cell assay (Table 1 ).
|
View this table: [in a new window] |
Table 1. Effects of M-CSF on NPA Secretion
|
We generated a SV40-transfected MEF cell line from the G–/–GM–/–/M–/– embryonic fibroblasts. To avoid the loss of potential M-CSF-dependent cells, cells were kept in M-CSF (5 ng/ml)-containing medium as soon as they were isolated from the embryos and during the transfection and selection steps. These cells (T-SV#22), immortalized and expanded in the constant presence of M-CSF, were then stimulated with LPS or HI C. albicans in the presence or absence of M-CSF (Fig. 5C) . M-CSF, added during the production of CM, enhanced NPA production significantly by immortalized fibroblasts, irrespective of the stimulus (Fig. 5C , P<0.05); however, once again, the NPA level was much lower than that obtained from the M-CSF+/+ cell line (SV#23). M-CSF alone, i.e., without addition of LPS or C. albicans, did not stimulate neutrophil clusters.
Role of M-CSF in the production of NPA by cells expressing the M-CSF gene product
The results obtained with G–/–GM–/–/M–/– embryos clearly show that M-CSF plays a critical role in NPA production. To confirm these results, we tested LPS-induced NPA production by immortalized G–/–GM–/–M+/+ cells (SV#23) in the presence of M-CSF-neutralizing antibody or with blocking anti-c-fms antibodies. Anti-M-CSF and anti-c-fms antibodies reduced NPA production significantly (P<0.05, Fig. 6
), although at the highest concentration tested, the anti-c-fms antibody was consistently less effective: This phenomenon may reflect ligand-independent but antibody-mediated aggregation and activation of c-fms. These antibodies were also tested against purified recombinant SCF, GM-CSF, and M-CSF to confirm their specificity: Both antibodies inhibited monocyte cluster formation by M-CSF, but neither affected the ability of SCF or GM-CSF to generate granulocyte clusters (data not shown); thus, the antibodies were specific for the M-CSF ligand or receptor. Antibodies to another surface molecule (cytokeratin-19) or growth factor (anti-SCF antibody) had no effect on NPA production (Fig. 6)
. The apparent inhibition of NPA production by the anticytokeratine antibody at the highest concentration tested was a result of cellular toxicity. Thus, specifically blocking M-CSF function during the production of the CM by G–/–GM–/– cells reduces NPA production to the levels observed in G–/–GM–/–M–/– cells.
 View larger version (15K): [in a new window] |
Figure 6. M-CSF is required for NPA production by M-CSF+/+ cells. SV#23 (Clone C4) cells were stimulated for 48 h with 0.05 µg/ml LPS in the presence of increasing amounts of neutralizing antibodies to M-CSF ( ), to the M-CSF receptor ( ), to SCF ( ), or to cytokeratine 19 ( ). The resulting CM were tested for NPA activity. The graph shows mean and SD of duplicate samples. Anti-M-CSF and anti-c-fms antibodies were tested six times with similar results.
|
 View larger version (46K): [in a new window] |
Figure 7. Differential secretion of cytokines by M-CSF+/– and M-CSF–/– cells. (A–C) Reactivity of CM with cytokine antibodies. Embryonic fibroblasts from G, GM–/–, M–/+ mice (A and B) or G, GM–/–, M–/– mice (C) were cultured in control medium (A) or medium supplemented with LPS (B and C). The resulting CM were applied to RayBio cytokine antibody Array II membranes and incubated overnight. The membranes were processed as described in Materials and Methods, and positive reactivity was detected using chemiluminescence. 1, Positive controls; 2, keratinocyte-derived chemokine (KC); 3, MCP-1; 4, RANTES; 5, soluble TNF receptor 1 (sTNFR1); 6, tissue inhibitor of matrix metalloproteinase 1 (TIMP-1); 7, eotaxin; 8, IL-6; 9, IL-12(p40/p70); 10, MIP-1; 11, MIP-2; 12, VEGF. (D) Template of the Cytokine Array II membranes. 6Ckine,; CTACK, cutaneous T cell-attracting chemokine; TARC, thymus and activation-regulated chemokine.
|
, MIP-2, and (weakly) VEGF. However, only MIP-1, IL-12(p40), and VEGF were absent in the CM from triple (G, GM, M-CSF)-KO fibroblasts, and the amount of IL-6 was reduced significantly. Thus, the presence or amounts of MIP-1, IL-12(p40), VEGF, and IL-6 correlated with NPA activity. We assessed the contribution to NPA activity of these cytokines by testing purified or commercially available recombinant proteins in the standard bone marrow assay. VEGF, MIP-1, and IL-12 were completely negative in the NPA assay, and IL-6 was partially active but reached plateau at levels well below those of NPA (Table 2
). Thus, the identity of NPA remains, for the moment, elusive. |
View this table: [in a new window] |
Table 2. NPA-Like Activity of Purified Cytokines
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In this study, we have used an in vitro system to help define the factor(s) that promote granulocyte generation in G-CSF/GM-CSF double-KO mice during infection. We have shown that embryonic fibroblasts from G-CSF/GM-CSF double-KO mice produce high levels of a NPA after stimulation with pathogens, such as HI C. albicans, or gram-negative bacterial components, such as LPS. The CM from these cells sustained a mixed population of early myeloid cells (as defined by Flt-3 positivity), which also expressed the myeloid markers Mac-1 and Gr-1. It is surprising that given the absence of G-CSF and GM-CSF, bone marrow progenitors exposed to NPA give rise to cells with the morphological, antigenic, and enzymatic characteristics of true granulocytes.
We have identified a crucial role for M-CSF in mediating LPS or C. albicans induction of NPA. Neither LPS nor C. albicans was able to elicit NPA production in cells derived from G–/–GM–/–/M–/– mice. Furthermore, neutralizing antibodies to M-CSF or the c-fms abolished NPA production from cells, expressing a functional M-CSF gene product; furthermore, addition of M-CSF to the MEF cultures partially restored NPA production. Our data therefore directly implicate M-CSF in NPA production by primary and immortalized MEFs and suggest that it plays an important regulatory role in granulocyte production in the G–/–GM–/– mice.
It is well recognized that M-CSF influences the proliferation and differentiation of hematopoietic stem cells along the monocyte/macrophage lineage and that M-CSF plays a critical role in LPS-induced macrophage activation [30
]. The macrophage lineage comprises a collection of cell types responsible for numerous metabolic, immunological, and inflammatory processes in physiological and pathological conditions [31
]. op/op mice [32
] are deficient in M-CSF and provide a unique model to understand the roles of M-CSF in development, physiology, and pathology. It has been shown that op/op mice lack a population of macrophages dependent on M-CSF, and the M-CSF-independent macrophage population, which resides in op/op mice, is primarily responsible for the classical macrophage functions in immunity, such as phagocytosis and antigen processing [15
]. The M-CSF-dependent macrophage population (absent in op/op mice) is primarily responsible for the regulatory functions of these cells, mediated by monokines. Thus, op/op mice have profound secondary deficiencies in cytokines secreted by this macrophage population, such as TNF-, IL-1, and G-CSF [33
, 34
]. In this context, it is not surprising that NPA secretion was also defective in mice null for M-CSF.
Despite a clear role for M-CSF in the production of NPA, the addition of exogenous M-CSF to G–/–GM–/–M–/– embryonic fibroblasts could only partially rescue NPA production, even when added to the cultures from the moment of MEF isolation. Other groups have shown that the deficiency of some tissue macrophage population in op/op mice is at most only partially corrected by injection of M-CSF [35 ]. Thus, it is likely that NPA is secreted by a M-CSF-responsive cell population, which is absent or reduced in the G–/–GM–/–M–/– cultures. We have not attempted in this study to identify or specifically select the cell population responsible for NPA secretion; however, the availability of our immortalized and clonally selected cell lines renders this identification possible in the future.
The identity of the cytokine(s) responsible for NPA activity remains to be determined: We have ruled out two of the most likely candidates, IL-3 and SCF, as well as other chemokines and cytokines such as IL-8, RANTES, and VEGF (Fig. 7 and data not shown). IL-6 may contribute to NPA activity but cannot replace NPA for the generation of granulocytes from bone marrow cells in mice lacking G-CSF and GM-CSF. Thus, granulopoiesis, particularly during exposure to pathogens, can be mediated by a variety of growth factors and cytokines with apparently redundant functions.
We suggest that production of NPA may play a role in emergency granulopoiesis in vivo and that M-CSF is pivotal in mediating this response. Of the cell types capable of mediating an immunological response to pathogens, the monocyte/macrophage lineage and the dendritic cell lineage, which share a a common precursor [36 ], are candidates for in vivo NPA-producing cells: Both secrete a plethora of cytokines and ILs [7 , 27 , 37 , 38 ], express c-fms [36 , 39 ], and are induced by M-CSF to differentiate and acquire functional capacity [40 , 41 ]. The important question is, of course, why there should be many cytokines with overlapping or redundant roles and what the function of NPA may be in an organism expressing functional G-CSF and GM-CSF genes. It is tempting to speculate that "NPA" and the classical granulocyte growth and differentiation factors (G-CSF and GM-CSF) may have differential temporal induction following exposure to pathogens, thus ensuring continuing and sustained production of neutrophils during infection. Unfortunately, a direct test of this hypothesis will have to await the identification of the cytokine(s) responsible for NPA and the direct measurement of its presence, activity, and biological relevance in normal mice.
Received January 11, 2007; revised May 30, 2007; accepted May 30, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
B-independent mechanism Br. J. Pharmacol. 131,1143-1153[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  F. Walker, H.-H. Zhang, V. Matthews, J. Weinstock, E. C. Nice, M. Ernst, S. Rose-John, and A. W. Burgess IL6/sIL6R complex contributes to emergency granulopoietic responses in G-CSF- and GM-CSF-deficient mice Blood, April 15, 2008; 111(8): 3978 - 3985. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
