| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Cooperative Research Centre for Chronic Inflammatory Diseases and Special Research Centre for Functional and Applied Genomics, Institute for Molecular Bioscience, University of Queensland, Brisbane, Queensland, Australia;
Cytopia Research Pty Ltd., Baker Heart Research Institute Building, Victoria, Australia; and
School of Molecular and Microbial Sciences, The University of Queensland, Brisbane, Queensland, Australia
1Correspondence: Institute for Molecular Bioscience, The University of Qld, St. Lucia, Qld, 4072, Australia. E-mail: m.sweet{at}imb.uq.edu.au
ABSTRACT
CSF-1 regulates macrophage differentiation, survival, and function, and is an attractive therapeutic target for chronic inflammation and malignant diseases. Here we describe the effects of a potent and selective inhibitor of CSF-1R—CYC10268—on CSF-1R-dependent signaling. In in vitro kinase assays, CYC10268 was active in the low nanomolar range and showed selectivity over other kinases such as Abl and Kit. CYC10268 blocked survival mediated by CSF-1R in primary murine bone marrow-derived macrophages (BMM) and in the factor-dependent cell line Ba/F3, in which the CSF-1R was ectopically expressed. CYC10268 also inhibited CSF-1 regulated signaling (Akt, ERK-1/2), gene expression (urokinase plasminogen activator, toll-like receptor 9, and apolipoprotein E), and priming of LPS-inducible cytokine production in BMM. In thioglycollate-elicited peritoneal macrophages (TEPM), which survive in the absence of exogenous CSF-1, CYC10268 impaired LPS-induced cytokine production and regulated expression of known CSF-1 target genes. These observations support the conclusion that TEPM are CSF-1 autocrine and that CSF-1 plays a central role in macrophage effector functions during inflammation. CSF-1R inhibitors such as CYC10268 provide a powerful tool to dissect the role of the CSF-1/CSF-1R signaling system in a range of biological systems and have potential for a number of therapeutic applications.—Irvine, K. M., Burns, C. J., Wilks, A. F., Su, S., Hume, D. A., Sweet, M. J. A CSF-1 receptor kinase inhibitor targets effector functions and inhibits pro-inflammatory cytokine production from murine macrophage populations.
Key Words: fms • imatinib mesylate • inflammation • BMM • TEPM
DIFFERENTIATION OF THE monocyte/macrophage lineage is primarily controlled by the growth factor CSF-1/M-CSF (1
2
3
4)
. CSF-1 acts through a single receptor, CSF-1R (5)
which is encoded by the c-fms proto-oncogene (6)
. Csf1op/Csf1op mice, which have an inactivating mutation in the CSF-1 gene, have gross deficiencies in macrophage numbers and effector functions (7)
, and gene targeting of the c-fms locus essentially resulted in a phenocopy of the Csf1op/Csf1op mouse (5)
. The CSF-1R is a member of the type III receptor tyrosine kinase family that includes other structurally related kinases such as Kit and platelet-derived growth factor (PDGF)-R (8)
. CSF-1 binding to the CSF-1R triggers dimerization, autophosphorylation, and activation of key signaling pathways including the PI-3 kinase/Akt and mitogen-activated protein kinase (MAPK) pathways. The gene expression program triggered by CSF-1 includes induction of genes involved in survival and proliferation such as Bcl-xL (9)
and cyclin D1 (10
, 11)
, as well as genes associated with mature macrophage effector functions like phagocytosis [e.g., scavenger receptor A (12)
and matrix remodeling [e.g., urokinase plasminogen activator (uPA) (13)
].
Since CSF-1 regulates key effector functions of macrophages, it plays important roles in a number of macrophage-mediated diseases. For example, CSF-1 potently primes macrophage responses to bacterial LPS (14
15
16)
. Hence, CSF-1 is likely to contribute to excessive inflammatory responses in Gram-negative sepsis. Excessive macrophage recruitment and survival have been linked to a number of chronic inflammatory states (17
18
19)
, and CSF-1 is likely to contribute to such processes. In a mouse model of rheumatoid arthritis—collagen-induced arthritis—administration of an anti-CSF-1 antibody (Ab) reduced disease severity whereas administration of exogenous CSF-1 exacerbated disease (20
, 21)
.
Blockade of CSF-1 action is also desirable for the treatment of malignant disease. The CSF-1R can be aberrantly expressed by tumors of epithelial origin such as breast and ovarian cancer (22
23
24)
, and expression of both CSF-1 and CSF-1R, as well as the CSF-1 target gene uPA, is strongly correlated with poor prognosis in these cancers (25
26
27)
. In murine models of breast cancer, constitutive CSF-1/CSF-1R signaling promoted metastatic disease, whereas interfering with this pathway reduced disease progression (28
29
30)
. Similarly, CSF-1 administration increased systemic vascular endothelial growth factor (VEGF) levels and promoted angiogenesis and solid tumor growth in mice (31)
.
Several lines of evidence suggest that targeting the CSF-1R is a viable clinical approach. The Abl/Kit/PDGF-R kinase inhibitor imatinib (Gleevec/STI571), which is widely used for treatment of malignant diseases such as chronic myeloid leukemia (CML), also inhibits activity of the CSF-1R (32
, 33)
. Another multitarget tyrosine kinase inhibitor (SU11248) inhibited CSF-1R-dependent osteolysis in a bone metastatic breast cancer model (34)
. A more selective CSF-1R inhibitor (GW2580) was recently reported and shown to inhibit CSF-1-dependent macrophage functions and tumor cell growth in vivo (35)
. In view of the many potential therapeutic applications of an inhibitor of CSF-1 action, we identified inhibitors of the CSF-1R kinase domain by screening a library of proprietary small molecule kinase inhibitors (36)
. In this study, we demonstrate the efficacy of CYC10268, a compound structurally distinct from GW2580 and SU11248, in regulating macrophage effector function.
MATERIALS AND METHODS
Cell culture and reagents
Cell lines and primary cell cultures were grown in RPMI 1640 medium containing 10% FCS, 20 U/ml penicillin, 20 µg/ml streptomycin and 2 mM L-glutamine in an incubator at 37°C with 5% CO2. The Ba/F3 cell line was maintained in medium containing 200 U/ml recombinant mouse interleukin (IL)-3 (a gift from Prof. Michael Waters, IMB, University of Queensland, Australia). BMM were derived from femurs of 6- to 8-wk-old, male inbred mice. Bone marrow cells were plated out in medium containing 104 U/ml (100 ng/ml) recombinant human CSF-1 (a gift from Chiron, Emeryville, CA, USA) on bacteriological plastic plates (Bibby Sterilin, Staffordshire, UK) for 7 days. Thioglycollate-elicited peritoneal macrophages (TEPM) were obtained by injecting BALB/c mice i.p. with 1 ml of 10% thioglycollate broth, followed by peritoneal lavage with 10 ml of PBS 5 days later. LPS from Salmonella minnesota (Sigma-Aldrich, St. Louis, MO, USA) was used at 10 ng/ml. IFN
(R&D Systems, Minneapolis, MN, USA) was used at 500 pg/ml. Kinase inhibitors were stored at 20 mM in DMSO at 4°C and DMSO provided the vehicle control in each experiment.
In vitro kinase assays
Tyrosine kinase assays were conducted using purified enzymes. Enzymes prepared as fusion proteins consisting of the kinase domain of the respective human kinase linked to glutathione S-transferase (GST) were purified by affinity chromatography. Kinase activities were determined using the amplified luminescence proximity homogeneous assay (ALPHA) screen (Perkin-Elmer, Wellesley, MA, USA) and biotinylated peptide substrates (Auspep Pty Ltd., Melbourne, Australia). For IC50 determinations, dose-response studies of putative kinase inhibitors were performed in 384-well Optiplates (Packard, Meriden, CT, USA) using a phosphotyrosine (P-Tyr-100) assay kit (Perkin-Elmer). Luminescence signals generated from the Optiplates were recorded in a Packard Fusion Alpha (Perkin-Elmer). Ten microliters of inhibitor diluted in assay buffer (10 mM HEPES, 25 mM NaCl, pH 7.5 containing 10 mM MgCl2, 0.01% Tween 20, 50 mM Na orthovanadate, and 0.1% BSA) was preincubated with 10 µl of the kinase in assay buffer for 20 min on ice. After centrifugation (1 min, 200 g), 10 µl of biotinylated peptide substrate (0.3 µM) containing 80 µM ATP in assay buffer was added and samples were incubated for 1 h at 30°C. After further centrifugation (1 min, 200 g), 10 µl each of streptavidin donor and antiphosphotyrosine acceptor beads diluted 1:100 in stop buffer (assay buffer containing 0.1M EDTA, 1% Tween-20, 0.1% BSA, pH 7.2) was added sequentially to each well, and the plate was incubated at room temperature for 1 h in the dark prior to reading luminescence in a Packard Fusion Alpha.
Plasmid construction and transfection
Full-length human c-fms cDNA obtained from Dr. S. Kellie (IMB, University of Qld, St. Lucia, Qld, Australia) was subcloned into pEF6/V5-His TOPO (Invitrogen, Carlsbad, CA, USA). Chimeric receptors were generated by splice overlap polymerase chain reaction (PCR) and cloned into pEF6/V5-His TOPO. All receptors were cloned in-frame with a V5-His C-terminal tag. The human CSF-1R extracellular domain was amplified with primers containing overlapping regions specific to the three intracellular domains of interest. The intracellular domains similarly overlapped the human extracellular domain, and the two domains were combined by performing PCR with both templates and the external primers. The mouse intracellular domain was amplified from a plasmid containing full-length mouse c-fms provided by T. Hamwood (IMB, University of Qld), while human flt3 and c-kit were amplified from THP-1 cDNA. PCR primers were as follows: human c-fms f: AGGCCATGGGCCCAGGA, human c-fms r: GCAGAACTGATAGTTGTTGG, human c-fms extracellular domain r: overlapping mouse c-fms: GCTTGTACTTGTACAATAGCAGCAGGAGC, overlapping human flt3: TTTTGTACTTGTACAATAGCAGCAGGAGC, overlapping human c-kit: TTTCTGTAAATATTTGTACAATAGCAGCAGGAGC; mouse c-fms intracellular domain f: GCTATTGTACAAGTACAAGCAGAAGCCGA, mouse c-fms intracellular domain r: GCAGAACTGGTAGTTGTTAG, flt3 intracellular domain f: GCTATTGTACAAGTACAAAAAGCAATTTAGGT, flt3 intracellular domain r: CGAATCTTCGACCTGAGC, c-kit intracellular domain f: CTGCTGCTATTGTACAAATATTTACAGAAACCCATGT, c-kit intracellular domain r: GACATCGTCGTGCACAAG. All clones were sequence verified. For generation of stable Ba/F3 transfectants, 5 x 106 cells were transfected by electroporation (1 pulse, 280V, 960 µF, GenePulser, Bio-Rad, Hercules, CA, USA) with 10 µg pEF6, which encodes Blasticidin resistance, in 250 µl complete medium + 1 mM HEPES. Stable transfectants were selected in 30 µg/ml Blasticidin (Invitrogen). Pooled stable transfectants were passaged into complete medium containing 104 U/ml CSF-1 and maintained in CSF-1-containing medium thereafter.
Cell viability assays
5 x 104 cells/well of a 96-well plate were plated in medium and treated for 24–48 h. For adherent cells (BMM and TEPM), growth medium was replaced with 1 mg/ml MTT in medium (Sigma Aldrich) and cells were incubated at 37°C for 1–2 h. MTT solution was then removed and the tetrazolium salt was solubilized with isopropanol for 10 min before the plates were read at 570 nm. For Ba/F3 cells, MTT stock solution (5 mg/ml in PBS) was added directly to growth medium at a concentration of 0.5 mg/ml and the plate was incubated at 37°C for 3–4 h. Solubilization was achieved with a solution of 10% SDS/50% isopropanol/0.01M HCl. The plates were read at 570 nm with a reference wavelength of 650 nm.
Apoptosis assays
For annexinV/propidium iodide staining, 1 x 106 Ba/F3-Human fms cells were plated in 6-well plates and treated with CSF-1, IL-3, and/or 3 µM CYC10268 for 48 h. AnnexinV/propidium iodide staining was performed using the annexin-V-Fluos Staining Kit (Roche Diagnostics, Penzberg, Germany) according to the manufacturer’s instructions. Cells were analyzed using a FACScalibur flow cytometer. For caspase 3/7 activation, 5 x 104 BMM were plated in 96-well plates and starved of CSF-1 overnight. CSF-1 and/or 3 µM CYC10268 were added the next morning and the assay was performed 24 h later using the Caspase-GloTM Kit (Promega, Madison, WI, USA) according to the manufacturers instructions. Luminescence was measured using a Packard TriLux Luminometer.
Measurement of proinflammatory cytokine release by ELISA
BMM or TEPM were plated in 24-well plates at 5 x 105 cells per well in 1 ml of complete medium with or without 104 U/ml CSF-1 (100 ng/ml) and/or 5 µM CYC10268 overnight. The next morning, cells were stimulated with 10 ng/ml LPS and supernatants were collected after 24 h and stored at –20°C until ELISAs were performed. ELISAs were conducted in triplicate using paired antibodies for IL-6, IL-12, and TNF
(BD PharMingen, San Diego, CA, USA).
Total RNA isolation and quantitative PCR
Total RNA was prepared from 3 x 106 cells using the Qiagen (Valencia, CA, USA) RNeasy Mini kit according to the manufacturers instructions. RNA was treated with DNase 1 (Ambion, Austin, TX, USA) and reverse transcribed using Superscript III (Invitrogen). cDNA levels of murine uPA, tlr9, ApoE, csf-1, c-fms, and hypoxanthine phosphoribosyl transferase (HPRT) were quantitated by real-time PCR using SYBR Green PCR master mix (Applied Biosystems, Warrington, UK) and an ABI Prism 7000 sequence detector (Applied Biosystems, Foster City, CA, USA). Amplification was achieved using an initial cycle of 50°C for 2 min and 95°C for 10 min, followed by 45 cycles of 95°C for 15 s and 50°C for 1 min. cDNA levels during the linear phase of amplification were normalized against HPRT controls. Primers (f: forward; r: reverse) used were as follows: upa f: GGTTCGCAGCCATCTACCAG, r: TTCCTTCTTTGGGAGTTGAATGAA; tlr9 f: AGGCTGTCAATGGCTCTCAGTT, r: TGAACGATTTCCAGTGGTACAAGT; csf-1 f: TCATGAGCAGGAGTATTGCCAA, r: GGCAATCTGGCATGAAGTCTC; c-fms f: CCACCATCCACTTGTATGTCAAAGAT, r: CTCAACCACTGTCACCTCCTGT; ApoE f: CTGGGTGCAGACGCTGTCT, r: CCTCCATCAGTGCCGTCAG; hprt f: GCAGTACAGCCCCAAAATGG, r: AACAAAGTCTGGCCTGTATCCAA.
Immunoblotting
BMM starved of CSF-1 overnight were treated with 104 U/ml CSF-1 in the presence or absence of 5 µM CYC10268. Whole-cell lysates were collected by lysing cell monolayers with boiling 66 mM Tris-Cl (pH 7.4)/2% SDS/1 mM sodium vanadate/1 mM sodium pyrophosphate/1 mM sodium moybdate/10 mM sodium fluoride. Extracts were resolved by SDS-PAGE, transferred to Imobilon-P (Millipore, North Ryde, NSW, Australia), blocked and probed with antiphospho protein antibodies [antiphospho-CSF-1R (Y721), antiphospho-p42/p44 MAPK, antiphospho-Akt (Cell Signaling Technology, Beverly, MA, USA)] in the presence of phosphatase inhibitors. Blots were washed, probed with HRP-labeled secondary antibodies (Cell Signaling Technology), and detected using enhanced chemiluminescence (ECL) reagents (Amersham Pharmacia Biotech, Piscataway, NJ, USA). Membranes were stripped with 66 mM Tris-Cl (pH 6.7)/2% SDS/100 mM 2-mercapto-ethanol and reprobed with total p42/44 MAPK or total Akt antibodies (Cell Signaling Technology).
Flow cytometry
CSF-1-starved Ba/F3 cells expressing human CSF-1R (or chimeric receptors) were washed and stained with a mouse monoclonal anti-human CSF-1R Ab (R&D Systems) or isotype control, followed by a PE-conjugated goat antimouse Ab (R&D Systems). Mouse CSF-1R was detected on BMM starved of CSF-1 overnight, as described previously (37)
. Cells were blocked in 0.1% goat serum and stained with a rat monoclonal antibody (mAb) against mouse CSF-1R (38)
, followed by a PE-conjugated goat anti-rat Ab (Serotec, Oxford, UK). Immunostained samples were detected using a FACSCalibur cytometer (Becton Dickinson, Franklin Lakes, NJ, USA).
IC50 calculation
The IC50s for CYC10268 and imatinib for each kinase domain in the Ba/F3 system were calculated from dose responses performed in quadruplicate over the dose range 0.001 µm to 100 µM. Data were analyzed using the GraphPad Prism software package. A nonlinear regression procedure was used to generate the curves from which IC50 values were determined. R2 (coefficient of determination) values for all curve fits were >0.92.
RESULTS
A structural analog of imatinib antagonizes CSF-1R in vitro kinase activity
The Abl tyrosine kinase inhibitor imatinib inhibits CSF-1R kinase activity (32)
, as well as a range of other tyrosine kinases (39
, 40)
. To identify small molecule inhibitors that would target CSF-1R but not Abl, a chemical library loosely based on the structure of imatinib was prepared and screened against recombinant human kinases, including CSF-1R and Abl (36)
. Several compounds inhibited activity of the human CSF-1R kinase in the low nM range (data not shown). One of these compounds, CYC10268, showed selectivity for the kinase domain of CSF-1R (Table 1
). CYC10268 did not inhibit the kinase activity of Flt3 that is closely related to CSF-1R or the activities of a range of other tyrosine kinases that are important regulators of macrophage function (Pyk2, Fes, Btk, Fak, Syk) (data not shown). For comparison, the effect of imatinib, which targets Abl and type III receptor tyrosine kinases, was also assessed. As expected, imatinib, but not CYC10268, inhibited the activity of Abl (Table 1)
. CYC10268 was 2-fold more potent on CSF-1R than imatinib and was 2- to 3-fold less potent on Kit and PDGF-R than imatinib (Table 1)
. Hence, CYC10268 is more selective for CSF-1R than imatinib and so was selected for analysis in functional assays of CSF-1 action.
|
CYC10268 inhibits ectopically expressed mouse and human CSF-1R and shows selectivity for CSF-1R kinase activity in cell culture systems
To determine whether CYC10268 could block CSF-1 signaling in a cell culture-based system, we stably transfected the factor-dependent cell line Ba/F3 with the human CSF-1R or a chimeric construct containing the human extracellular CSF-1R region and the intracellular domain of murine CSF-1R. Use of a chimeric construct ensured that both transfected lines have the same extracellular domain for ligand recognition, thus enabling a more reliable comparison of human and mouse CSF-1R kinase activities in this system. Pooled stable clones were selected for survival in CSF-1. Parental Ba/F3 cells do not express CSF-1R and do not respond to CSF-1 (41)
. Both stably transfected Ba/F3 cell populations expressed CSF-1R on the cell surface (Fig. 1
A) and proliferated in response to CSF-1 (Fig. 1B, C
). CYC10268 (1 µM) inhibited survival of CSF-1R-positive Ba/F3 cells in response to CSF-1, but not IL-3, displaying similar potency with respect to the mouse and human kinases (Fig. 1B, C
). This effect correlated with the ability of CYC10268 to up-regulate cell surface phosphatidyl serine, an event associated with the onset of apoptosis, in cells growing in CSF-1, but not IL-3 (Fig. 1D
).
|
To determine whether the selectivity of the inhibitors observed in in vitro kinase assays was also apparent in this cellular system, Ba/F3 cells expressing chimeric receptors containing the extracellular domain of human CSF-1R and the intracellular domain of human Kit or Flt3 were generated. Both lines expressed high levels of surface CSF-1R (Fig. 2
A) and proliferated in response to CSF-1 (data not shown). CYC10268 did not affect survival mediated by the Flt3 kinase domain except at high concentrations (Fig. 2B
). As is the case with imatinib, CYC10268 inhibited survival mediated by the Kit kinase domain (Fig. 2B
). We therefore directly compared the actions of CYC10268 and imatinib on the kinase activity of CSF-1R and Kit in this system (Table 2
). CYC10268 was 
2- to 3-fold more potent than imatinib at preventing survival mediated by the human and mouse CSF-1R kinase domains. Further, whereas CYC10268 affected CSF-1R and Kit-mediated survival with similar potency, imatinib was 5-fold more potent on Kit than on human or mouse CSF-1R. Hence, although CYC10268 is not completely specific for the CSF-1R, its selectivity for CSF-1R was enhanced compared to imatinib.
|
|
CYC10268 blocks CSF-1-induced signaling and survival of bone marrow-derived macrophages (BMM)
Bone marrow-derived macrophages, which are routinely used to study CSF-1-mediated signaling and gene regulation, undergo apoptosis when starved of CSF-1 (42)
. Apoptosis can be rescued by either CSF-1 or bacterial products that signal through the CSF-1R and toll-like receptors (TLRs), respectively, to down-regulate caspase 3/7 activation (43)
and induce prosurvival genes such as Bcl-xL (44)
. LPS also induced expression of csf-1 mRNA in keratinocytes and bone marrow stromal cells (45
, 46)
, raising the possibility that LPS-initiated survival in macrophages is mediated by autocrine CSF-1. We therefore assessed the ability of CYC10268 to prevent CSF-1- and LPS-induced survival of BMM. CYC10268 completely blocked CSF-1-induced survival but had little effect on LPS-induced survival (Fig. 3
A). Similarly, CYC10268 prevented suppression of caspase 3/7 activity by CSF-1, but not LPS (Fig. 3B
). Hence, LPS-mediated survival of BMM was not mediated by autocrine CSF-1. CYC10268 blocked CSF-1-mediated survival whether administered before CSF-1 or to CSF-1-replete BMM, indicating that it targeted constitutive CSF-1 signaling. The failure of CYC10268 to affect LPS-mediated survival in the 1–10 µM range also confirms that its action on macrophage survival was not via a general cytotoxic effect. The morphology of LPS-treated BMM was not affected by CYC10268, whereas there was a clear effect on the morphology of CSF-1-treated BMM (Fig. 4
).
|
|
CSF-1 binding to the CSF-1R triggers receptor autophosphorylation, internalization, and lysosomal degradation of the CSF-1R (47)
. CYC10268 inhibited CSF-1R kinase activity as measured by receptor autophosphorylation on Y721 (Fig. 5
A). We next assessed whether CYC10268 affected receptor endocytosis since there are conflicting studies of the importance of CSF-1R tyrosine kinase activity in this process (48
49
50
51)
. Treatment of CSF-1-starved BMM with CSF-1 for 20 min down-regulated cell surface CSF-1R (Fig. 5B
), as we and others have shown (37
, 52
, 53)
. CYC10268 did not affect CSF-1R down-modulation in response to CSF-1 (Fig. 5B
). Hence, this event is not dependent on CSF-1R kinase activity.
|
To confirm that CYC10268 blocked signaling downstream of the CSF-1R, we examined CSF-1-induced ERK1/2 and Akt phosphorylation (10
, 54
, 55)
. At concentrations that inhibited CSF-1-mediated survival, the inhibitor blocked phosphorylation of Akt and ERK-1/2 in response to CSF-1 (Fig. 5C
). We also examined the effect of CYC10268 on CSF-1- and LPS-induced phospho-tyrosine profiles in BMM. Treatment of BMM with CSF-1 or LPS for 8 h triggered tyrosine phosphorylation, but only the CSF-1 response was inhibited by CYC10268 (Fig. 5D
).
Modulation of macrophage gene expression and activation by CYC10268
CSF-1 regulates expression of genes involved in macrophage effector functions. For example, CSF-1 triggers macrophage migration that is partly mediated by up-regulation of genes involved in degradation of the extracellular matrix (ECM) such as uPA (13)
. CSF-1 also down-regulates expression of tlr9 mRNA, which is required for responses to CpG DNA (16)
. We have also shown by microarray that CSF-1 dramatically suppresses apolipoprotein E (ApoE) mRNA expression (unpublished data). We therefore assessed whether CYC10268 blocked the ability of CSF-1 to regulate expression of uPA, tlr9, and ApoE mRNAs. CYC10268 had marginal effects on basal uPA, tlr9, and ApoE gene expression in CSF-1-starved BMM. These effects (repression of uPA mRNA expression and modest induction of ApoE and tlr9 mRNA expression) are likely to reflect inhibition of low basal CSF-1R activity by CYC10268 in CSF-1-starved cells. CYC10268 blocked the induction of uPA mRNA and repression of tlr9 and ApoE mRNAs in response to CSF-1 (Fig. 6
A).
|
Another major effector function regulated by CSF-1 is macrophage proinflammatory cytokine production. CSF-1 potently primes macrophage responses to LPS (14
15
16
, 56)
. We therefore assessed whether CYC10268 prevented CSF-1 from priming the macrophage response to LPS (Fig. 6B
). Overnight treatment of BMM with CSF-1 enhanced LPS-induced TNF-, IL-6, and IL-12 production 5- to 50-fold and treatment with CYC10268 blocked the ability of CSF-1 to prime proinflammatory cytokine production. IFN also primes macrophage responses to LPS, but CYC10268 did not inhibit this priming effect.
CYC10268 affects the function of thioglycollate-elicited peritoneal macrophages (TEPM)
Unlike BMM, TEPM are not dependent on exogenous CSF-1 for survival. This may be because TEPM produce low levels of endogenous CSF-1 or receive an alternative survival signal. Hence, we assessed the effect of CYC10268 on survival and expression of CSF-1-responsive genes in TEPM. CYC10268 at doses of 1–5 µM (which did not cause toxicity in BMM cultured without CSF-1) (Fig. 3A
) reduced the viability of TEPM cultured in the absence of CSF-1 by 2-fold (Fig. 7
A). Although CSF-1 is not required for TEPM survival, TEPM do express the CSF-1R and respond to CSF-1 (57
58
59
60)
. CSF-1 did enhance the viability of TEPM and this effect was ablated by treatment with 5 µM inhibitor (Fig. 7A
). The ability of the CSF-1R kinase inhibitor to at least partially block TEPM viability suggests that TEPM do receive a survival signal through the CSF-1R. We therefore assessed csf-1 mRNA expression in TEPM vs. BMM. Figure 7B
indicates that TEPM expressed 17-fold higher levels of csf-1 mRNA than BMM. These levels were comparable to those observed in primary osteoblasts (Dr. Liza Raggatt, personal communication) that produce CSF-1 at biologically active levels. Such an effect would account for constitutive uPA expression in TEPM. We therefore assessed whether basal uPA mRNA expression in TEPM was affected by CSF-1R kinase inhibitors. Figure 7C
demonstrates that CYC10268 suppressed basal uPA expression over a 24 h time course. CSF-1 had a minimal effect on uPA mRNA levels in TEPM, and CYC10268 maintained low basal uPA expression in the presence of exogenous CSF-1 (Fig. 7C
).
|
These data indicate that TEPM essentially behave as CSF-1-primed macrophages. Since CSF-1 primes murine macrophage populations for production of proinflammatory cytokines in response to LPS (15)
, we hypothesized that CYC10268 would impair the release of proinflammatory cytokines from TEPM in response to LPS. TEPM were treated overnight with or without CYC10268, then stimulated the next day with LPS. CYC10268 dramatically impaired LPS-induced production of TNF-, IL-6, and IL-12 from TEPM (5-fold, 10-fold, 10-fold, respectively) irrespective of CSF-1 pretreatment. The effect of CYC10268 on LPS-induced proinflammatory cytokine production is unlikely to be a consequence of decreased survival, as survival was only modestly affected by comparison (Fig. 7A
). Hence, targeting the CSF-1 signaling pathway severely impairs production of inflammatory mediators from murine macrophage populations.
DISCUSSION
Dissecting the role of CSF-1 in the effector function of mature macrophages and in inflammatory disease is difficult because of the absolute requirement for CSF-1 in normal macrophage development. Deficiencies in macrophage development are likely to be major contributors to dysregulated inflammatory responses in CSF-1 or CSF-1R deficient mice (3
, 5
, 61
, 62)
. Consequently, the role CSF-1 plays in regulating effector functions of mature macrophages during inflammatory responses has been difficult to assess (18)
. Here we introduce a potent CSF-1R kinase inhibitor that will provide further insight into the role of the CSF-1/CSF-1R signaling system in mature macrophage effector function. The potency of CYC10268 was demonstrated in a range of biological assays and it targeted CSF-1 signaling in macrophages that were pretreated with CSF-1 several hours before the addition of inhibitor (Fig. 3A, B
). Since CYC10268 targeted both the human and mouse CSF-1R kinase domains (Fig. 1B, C
), this compound has potential in both experimental animal models of inflammatory and metastatic disease, as well as in clinical applications.
CYC10268 could be trialed in a range of diseases where CSF-1 has been implicated. These include chronic inflammatory diseases such as rheumatoid arthritis (20)
, inflammatory bowel disease (63)
, kidney disease (64
, 65)
, and cardiovascular disease (66)
. Indeed, CYC10268 blocked macrophage proliferation and CSF-1-dependent priming for production of proinflammatory cytokines that are associated with chronic inflammation (Fig. 3A
, Fig. 6B
, Fig. 7A, D
). The clear link between CSF-1/CSF-1R and metastatic potential (23
,24
,67)
is primarily thought to involve expression of enzymes such as uPA that degrade the ECM (22)
. Since CYC10268 blocked CSF-1-induced expression of uPA (Fig. 6A
, Fig. 7C
), it could also have potential in malignant disease. CYC10268 might also be utilized for enhancing vaccine efficacy. CSF-1 is immunosuppressive for T cell responses in vitro (68)
and in vivo (69)
, and a recent report showed that imatinib enhanced antigen-presenting cell function (70)
. In addition, CSF-1 repressed expression of TLR9 (Fig. 6A
) (16)
, which is required for responses to CpG DNA (71)
. CpG DNA has been used widely as a vaccine adjuvant (72
73
74)
. Hence, targeting CSF-1 during vaccine delivery may enhance efficacy by promoting antigen presentation as well as increasing the adjuvant activity of CpG DNA.
CYC10268 was selective for CSF-1R over a number of other kinases in in vitro kinase assays (Table 1)
. Selectivity was also assessed in cell culture systems. When used at levels that abrogated CSF-1R-dependent survival, the compound did not inhibit survival mediated by the highly related Flt3 kinase in Ba/F3 cells (Fig. 2B
, Table 2
). CYC10268 was not, however, completely specific for CSF-1R since survival mediated by the Kit kinase domain was impaired in Ba/F3 cells (Fig. 2B
, Table 2
). It should be noted that there were some differences in inhibitor potency between enzyme (Table 1)
and cell-based (Table 2)
assays. In particular, the activities of CYC10268 and imatinib on Kit relative to CSF-1R were enhanced in the Ba/F3 system. This highlights the need to perform inhibitor selectivity studies in a cell-based system, as well as in recombinant enzyme assays.
Apart from CYC10268, imatinib (32
,33)
, SU11248 (34)
, and GW2580 (35)
all inhibit CSF-1R kinase activity. However, each inhibitor has distinct specificities for tyrosine kinases. GW2580 appears to be relatively selective for CSF-1R (35)
; imatinib targets Abl/Kit/PDGFR as well as CSF-1R (Tables 1
and 2)
(32
,39
,75)
; SU11248 targets a broad range of tyrosine kinases including VEGFR, PDGFR, Kit, Flt3, and CSF-1R (76)
; CYC10268 targets CSF-1R, Kit, and PDGFR (Table 1)
. There are also differences in the potencies of the inhibitors on the kinases they do target. For example, while CYC10268 and imatinib both target CSF-1R and Kit, CYC10268 was 2- to 3-fold more potent than imatinib on CSF-1R (Table 1
, Table 2
). Conversely, imatinib was more potent than CYC10268 on Kit (Tables 1
, 2)
.
There are several reasons why alternatives to imatinib might be pursued for therapeutic applications, and each of the compounds described above might have distinct advantages in different therapeutic applications. First, imatinib has only been used clinically in the last 5 years. Therefore, its long-term effects are unknown. Second, although imatinib is well tolerated in patients, a variety of hematologic and nonhematologic side effects have been reported (76
, 77)
. Apart from its myelosuppressive effects, imatinib also directly inhibited T cell proliferation, implying that it might adversely affect immune responses (78)
. Hence, there is scope for the trial of other inhibitors of CSF-1 action with the goal of reducing such side effects. Third, apart from its use in the treatment of CML, imatinib is used to treat gastrointestinal stromal tumors (GISTs) in which Kit or PDGFR function is dysregulated (79
, 80)
. CYC10268 may have therapeutic potential in the treatment of GISTs, whereas GW2580, which does not target Kit or PDGFR, is unlikely to be effective. Finally, imatinib resistance can occur in both CML and GISTs, and alternative compounds might be beneficial. Indeed, SU11248 overcame imatinib resistance for a mutant form of Kit (81)
. CYC10268 could have similar applications.
CYC10268 also provides a powerful tool to dissect the role of CSF-1/CSF-1R signaling in mature macrophage effector functions. In this study, we have used CYC10268 to assess several aspects of CSF-1 signaling. First, we show that CSF-1-dependent CSF-1R internalization is not dependent on CSF-1R kinase activity (Fig. 5B
). We also show that LPS-mediated survival of BMM does not occur via a CSF-1-dependent pathway (Fig. 3A
). Hence, although LPS can induce CSF-1 expression in certain cell populations (45
, 46)
and in vivo (82)
, it acts independently of CSF-1 to trigger survival in BMM. We also show that TEPM, which do not require exogenous CSF-1 for survival, behave as CSF-1 primed cells. TEPM are frequently used as a model for inflammatory macrophages but are, in fact, a relatively ill-defined population. TEPM proliferation and colony-forming capacity in the presence of TEPM-conditioned medium, prior to the identification of CSF-1, led to suggestions that TEPM cultures were autocrine for an essential growth factor (57)
. Autocrine TEPM survival is unlikely to be mediated by stem cell factor since its receptor, Kit is down-regulated during macrophage differentiation (38)
. We show here that TEPM constitutively express both c-fms and csf-1 mRNAs (Fig. 7B
), and treatment of TEPM with CYC10268 regulated the basal expression of the CSF-1 target gene, uPA (Fig. 7C
). Furthermore, TEPM produced high levels of proinflammatory cytokines in response to LPS. CSF-1 did not enhance this response, but CYC10268 potently inhibited LPS-induced cytokine production (Fig. 7D
). This supports the notion that constitutive CSF-1 "priming " is part of the TEPM inflammatory phenotype, and that blockade of CSF-1 signaling is likely to be effective in suppressing macrophage effector functions.
In summary, we have clearly demonstrated the potency and selectivity of a novel inhibitory compound that targets the CSF-1R. This compound and its derivatives will be useful for understanding the biology of CSF-1 action and should have a broad range of potential therapeutic applications. A comparison of the efficacies and tolerance of the different CSF-1R kinase inhibitors in different models of inflammatory and metastatic disease should be a focus of future research.
ACKNOWLEDGMENTS
We thank Dr. Emmanuelle Fantino and Ms. Margarita Kurek for enzymes and enzyme assays, respectively, and Dr. Michael Harte and Dr. Xianyong Bu for preparation of CYC10268. We also wish to thank Professor Mike Waters and Dr. Richard Brown for providing Ba/F3 cells and recombinant IL-3. This work was supported by Cytopia Pty Ltd. under an ARC Linkage Grant (LP0454363); and by NHMRC project grant 301210. Several authors (C.B., S.S., A.W.) have declared a financial interest in a company (Cytopia) whose potential product was studied in the present work. Several (C.B., S.S., A.W.) are employed by a company (Cytopia) whose potential product was studied in the present work. C.B. and A.W. hold a patent related to the work described in the present study.
Received for publication February 8, 2006. Accepted for publication April 17, 2006.
REFERENCES
