Neutropenia with Impaired Immune Response to Streptococcus pneumoniae in Ceramide Kinase-Deficient Mice

Ceramide kinase (CerK)² uses ATP to phosphorylate ceramide, forming ceramide-1-phosphate (C1P) (1). In mammals, this is the only known pathway leading to the generation of C1P since alternative putative pathways have not passed validation assays. First, we recently characterized a CerK-like protein and variants thereof. Although the closest homolog known for CerK, all CerK-like proteins failed to phosphorylate ceramide in our assays (2). Second, a more distant homolog named multisubstrate lipid kinase was originally described as capable of using ceramide as a substrate in vitro (3). However, this has been challenged (4) and the enzyme has recently been renamed acylglycerol kinase (5).

Based on experiments with immortalized cell lines, the use of RNA interference or of a recently reported inhibitor, CerK and C1P have been suggested to play a role in inflammation by controlling mechanisms such as phagocytosis (6, 7), mast cell degranulation (8, 9), and cytosolic phospholipase A₂α (cPLA₂α)-dependent eicosanoid synthesis (10, 11). Therefore, defining how CerK and C1P may impact on innate immunity is of particular interest. To this aim, we generated CerK-deficient mice (Cerk^−/−) by homologous recombination and we report here on their broad characterization, with particular emphasis on responses following challenges of the innate branch of the immune system.

Cerk^−/− mice were tested in a variety of disease settings dependent on macrophage responses such as the Ag-induced (12) and serum transfer models (13) of rheumatoid arthritis. In bone marrow-derived macrophages (BMDM), we investigated phagocytosis and the capacity for angiogenic switch, as well as pathways where cPLA₂α was previously reported to be essential (14, 15). Mast cell function was probed in allergic disease models, as well as using a panel of in vitro readouts following IgE/Ag stimulation, e.g., arachidonic acid and eicosanoid release, degranulation, cytokine release, and calcium flux. Finally, neutrophils, which are major players for innate defense against a variety of pathogens, were investigated using a model of pulmonary infection.

PMA, LPS 0111:B4 from Escherichia coli, and ionomycin were obtained from Sigma-Aldrich. A23487, geneticin, and N-[7-(4-nitrobenzo-2-oxa-1,3-diazole)] (NBD)-C6-ceramide were purchased from Invitrogen Life Technologies. Efipladib (Wyeth-1) was synthesized at the National Institutes for BioMedical Research. All other reagents were purchased from Sigma-Aldrich unless otherwise stated.

CerK genomic sequences corresponding to CerK introns 1 and 5 were amplified from BALB/c mouse genomic DNA and subcloned into vector pRAY2 (accession no. U63120), resulting in the targeting plasmid for homologous recombination (pCerKtarget). Subcloned sequences were compared with sequences available from the Ensembl database (Ensembl Gene Identification ENSMUSG00000035891). BALB/c mouse embryonic stem cell culture was performed with primary x-ray-inactivated embryonic fibroblasts derived from DR4 mice. Embryonic stem cells were transfected by electroporation using 20 μg of NotI-digested pCerK. Transfected cells were selected for neomycin resistance using 0.2 mg/ml geneticin. Ten days after transfection, 400 G418-resistant cell clones were isolated and analyzed by PCR for homologous recombination. For PCR genotyping of selected embryonic stem cells, cells were extracted in 50 μl of lysis buffer (10 mM Tris-HCl (pH 8.0), 0.05% SDS, and 50 μg/ml proteinase K), and PCR was performed using 1 μl of crude cell extract in a total volume of 20 μl. PCR were performed using the Qiagen TaqPCR MasterMix. For the first PCR, 1 μl of genomic DNA was mixed with 0.8 μl of a 10 μM solution of sense primer (TCCTCGTGCTTTACGGTA) and a 10 μM solution of antisense primer (TCCTCTTCTTGCAGTAGCTG). Ten microliters of Qiagen TaqPCR MasterMix as well as 7.4 μl of H₂O were added; amplification was for 35 cycles (94°C for 30 s, 55°C for 30 s, and 72°C for 2 min). One microliter of the first reaction was taken for nested PCR using the same conditions and the sense primer CTATCGCCTTCTTGACGA and antisense primer CAAGATTGGCCTACCAGACA. For Southern blot, 5 μg of genomic DNA was digested with 30 U of the restriction enzyme (as indicated in the legend to Fig. 1) and separated on a 0.9% agarose gel. After denaturation, the DNA was blotted on a Hybond N⁺ membrane (GE Healthcare) followed by UV cross-linking. Hybridization with the ³²P-labeled neomycin DNA probe (Rediprime II Random prime labeling kit; GE Healthcare) was performed in Perfect Plus Hybridization buffer (Sigma-Aldrich) at 65°C overnight. After washing the hybridized membrane, image analysis was performed using a STORM 860 PhosphorImager (Molecular Dynamics). Targeted BALB/c embryonic stem cells were injected into C57BL/6 host blastocysts and transferred into pseudo-pregnant CB6F₁ foster mothers. Chimeric offspring were identified by coat pigmentation (white (BALB/c) on a black (C57BL/6) background). White offspring indicated the germline transmission of the targeted embryonic stem cells and were further analyzed for their correct genotype by PCR. Genotyping was performed on tail-biopsy genomic DNA using a Sigma-Aldrich REDExtract-N-Amp Tissue PCR kit and the following primers: sense, TCCTCGT GCTTTACGGTA, and antisense, TCCTCTTCTTGCAGTAGCTG, to amplify a 1.7-kb fragment of the Cerk^−/− gene locus; sense, AGAATGGATCCTGAGATGGC, and antisense, as above, to amplify a 1.7-kb fragment of the WT Cerk locus. After 2 min at 94°C (hot start), 35 cycles were performed in the following conditions: 94°C for 30 s, 56°C for 30 s, and 68°C for 1 min 30 s (10 min for the 35th cycle). The animals were kept under standard housing conditions. All investigations were performed on 6- to 12-wk-old mice. The experimental procedures met all regulations and standards as approved by the Austrian and Swiss governments.

Total RNA was isolated using an Absolutely RNA Miniprep Kit (Stratagene). For real-time PCR, an Applied Biosystems PRISM 7900HT Sequence Detection System was used. Primers were designed using Applied Biosystems Primer Express 2.x software (2). For relative quantification, data were analyzed using the ΔΔCT method as described previously (16). Expression levels of target genes in each sample were normalized to the average of housekeeping genes.

Cells (1–2 × 10⁶/sample) were lysed by freeze/thawing in 400 μl of PBS. Two hundred microliters was transferred into a glass tube and 300 μl of 2 M KCl/4% HCl as well as 1875 μl of methanol/chloroform 2:1 were added. Samples were spiked with an internal standard (C8-C1P; Avanti Polar Lipids). After vortexing and sonication, 675 μl of chloroform was added and, after further vortexing, 500 μl of water. The sample was finally centrifuged for 10 min at 2600 rpm at 4°C. The organic phase was dried in an autosampler vial. Samples were dissolved in mobile phase and subjected to HPLC (HPLC 1100; Agilent), performed as previously described (17) with some modifications. In short, a Supelco Discovery C₁₈ HS column (2.1 mm × 50 mm, 3-μm particle size) was used and eluted with a gradient (eluent A: 5 mM ammonium formiate plus 1% formic acid in methanol:water 70:30 plus 2% tetrahydrofuran; eluent B: 5 mM ammonium formiate plus 1% formic acid in methanol plus 2% tetrahydrofuran; 70–100% B in 1.8 min) at a flow of 400 μl/min at 60°C. Electrospray-ionization with tandem mass spectroscopy (MS/MS) using an API 4000 QTrap instrument (MDS Sciex) was used to detect C1P with positive ionization. The optimal collision energy for C16-C1P was +51 V and for C8-C1P was +39 V; the multiple reaction monitoring transitions followed were m/z 618.6/264.1 and 506.4/264.2, respectively.

Cells were seeded to a 6-well plate. On day 7, cells were washed and preincubated for 6 h in phosphate-free medium (MEM Earl) supplemented with 10% FCS (dialyzed against TBS). Then medium was replaced by medium containing 100 μCi/ml [³³P]orthophosphate (GE Healthcare). After 15 h of incubation, the medium was removed and cells were washed twice with PBS, scraped with 0.5 ml of cold methanol, and transferred to Eppendorf vials. After addition of 0.5 ml of chloroform and 0.45 ml of 0.5 N HCl/2 M NaCl, samples were vortexed and after a short centrifugation the chloroform phase was analyzed by two-dimensional TLC (Silica 60 HPTLC plates; Merck) with chloroform:methanol:25% NH₄OH (60:35:8) for the first dimension and chloroform:methanol:acetone:acetic acid:water (10:4:3:2:1) for the second dimension as previously described (4). Phospholipids were visualized by autoradiography. For NBD-C6-ceramide-based assays, NBD-C6-ceramide was added to cells at 5 μM in cell culture medium and allowed to incubate for 2 h. Lipids were extracted and run on TLC plates using butanol:acetic acid:water (3:1:1) as mobile phase.

Ceramide concentrations in serum were determined by HPLC (Agilent 1100 chromatograph) with detection by electrospray-ionization MS (API 4000 QTrap instrument; MDS Sciex) with Multiple Reaction Monitoring (MRM). The procedure followed published methods (17, 18), with some modifications. Aliquots of serum (100 μl) were spiked with internal standard (C17-ceramide, final concentration 1 ng/ml; Avanti Polar Lipids) and 750 μl of methanol:chloroform (2:1) was added. After mixing and incubating for 1 h at 48°C, the mixture was cooled to room temperature and 75 μl of 1 M KOH was added, followed by a 1- h incubation at 37°C. The pH was then adjusted to ∼5 by addition of 10% acetic acid (∼100 μl). After addition of 750 μl of chloroform, the samples were mixed for 3 min and centrifuged (2600 rpm, 10 min). The organic layer was dried in vacuo and reconstituted in an eluent mixture of 80% B and 20% A (see below). Samples were chromatographed on a Luna C₁₈ column (3 μm, 2 × 50 mm; Phenomenex), which was eluted with a gradient (eluent A: 5 mM ammonium acetate in acetonitrile containing 2% methanol, 1% acetic acid, 0.25% tetrahydrofuran; eluent B: 5 mM ammonium acetate in methanol containing 1% acetic acid and 0.25% tetrahydrofuran). The gradient was shaped as follows: 4 min at 1% B isocratic, then in 6 min to 100% B at a flow rate of 0.6 ml/min and a column temperature of 50°C. The MRM transitions monitored (m/z) and the optimal collision energies in [V] were as follow [ceramide acyl chain length/Q1 MRM transition; Q3 MRM transition/optimal collision energy]: [C12/482.39;264.5/33] [C16/538.44;264.5/33] [C16-Dh/540.46;284.5/39] [C17/552.41;264.5/41] [C18/566.43;264.5/35] [C18-Dh/568.51;284.5/33] [C18:1/564.48;264.5/35] [C20/594.47;264.5/37] [C24/650.54;264.5/41] [C24-Dh/652.54;284.6/37] [C24:1/648.54;264.5/45] [C24:1-Dh/650.53;284.5/41]. For calibration, serum samples were spiked with 0.05–2500 ng/ml of the individual ceramides added from stock solutions in methanol:chloroform (9:1). Analytical standards were all bought from Avanti Polar Lipids.

Heparinized blood samples were collected from the orbital sinus. After determination of the total number of leukocytes with a hemocytometer (Coulter AcT diff), differential cell counts of stained blood smears were performed based on morphology and staining characteristics. The absolute lymphocyte counts were calculated by multiplying these percentages by the total number of white blood cells.

Single splenocyte suspensions were prepared in PBS by gently pressing the spleens through a 70-μm nylon cell strainer (BD Falcon). Total white blood cell counts in splenocyte suspensions, blood, and lavage fluids were determined with a cell counter (Coulter AcT diff). For bone marrow preparations, the femurs from hind legs were removed and PBS solution was forced through the bone with a syringe. After dispersing cell clumps, the cell suspension was centrifuged (400 × g, 10 min, 4°C) and resuspended in PBS. FACS analysis was performed using rat anti-mouse Gr1 (Ly6G/Ly6C; clone RB6-8C5) and rat anti-mouse CD11b mAbs (BD Pharmingen). Neutrophils were identified as Gr1^highCD11b⁺ (R1) as opposed to monocytes that are Gr1^lowCD11b⁺ (R2).

BMMC from WT and Cerk^−/− mice were cultured as previously described (19) in medium supplemented with 100 ng of stem cell factor and 10 ng of IL-3/ml. Differentiation of mast cells was monitored as previously described (19). Cells were used when >95% of the population expressed Fcε RI. In brief, 1 × 10⁶ cells/ml were sensitized with 2 μg of murine IgE/ml and stimulated with 100 ng of DNP-BSA/ml (Calbiochem).

Cells were challenged with Ag/IgE for 1 h (for degranulation and arachidonic acid release measurements, see Fig. 5) or for 4 h (for TNF-α release measurement, see Fig. 5, and MCP-1 release measurement, data not shown), and degranulation was determined by measuring the enzymatic activity of the granule marker β-hexosaminidase as previously described (19). The release of leukotrienes C4/D4/E4 (enzyme immunoassay from GE Healthcare), PGD₂ (enzyme immunoassay from Cayman Chemical), and cytokines (BD OptEIA mouse TNF-α ELISA or mouse MCP1-ELISA from BD Biosciences) into the supernatant were measured according to the manufacturers’ protocols. Arachidonic acid release was measured as described below for macrophages.

Cells were loaded with 5 μM Fluo-4/AM in HBSS supplemented with 2.5 mM probenecid for 1 h and then washed twice with HBSS. Cell suspensions were plated onto 96-well plates at a density of 2 × 10⁵ in 50 μl of medium/well and then analyzed for calcium mobilization using a FlexStation (Molecular Devices) with an excitation wavelength at 485 nm and an emission wavelength at 530 nm.

One milliliter of Brewer’s modified thioglycolate medium (BD Pharmingen) was injected i.p. into mice. Three days later, mice were sacrificed and peritoneal cavities were lavaged twice, each time with 2.5 ml of ice-cold Ca²⁺- and Mg²⁺-free PBS (PBS⁻). Intraperitoneal cells were pooled, washed twice with ice-cold PBS⁻, and cultivated for 2 days at 37°C in a humidified atmosphere containing 95% air and 5% CO₂ to remove nonattached lymphocytes, neutrophils, and mast cells. Cells were washed twice with warm PBS⁻ and macrophages attached onto the culture dish were removed with 1 ml of 0.25% trypsin and 1 mM EDTA solution, then seeded for experiments.

Peritoneal macrophages, 2 × 10⁵ cells/well in a 24-well plate in RPMI 1640 medium containing 10% heat-inactivated FCS, were loaded with 5 nM (1 μCi/ml) tritiated arachidonic acid ([5,6,8,9,11,12,14,15-³H]arachidonic acid; GE Healthcare) in medium and incubated at 37°C for 16 h. Cells were washed twice and stimulated according to the legend of Fig. 3. Cell culture supernatants were harvested and, after centrifugation, an aliquot was used for scintillation counting. PGE₂ was determined in supernatants of unlabeled cultures using the PGE₂ enzyme immunoassay from Cayman Chemical. For leukotriene release, resident peritoneal macrophages from WT and Cerk^−/− mice were seeded into a 24-well plate with 1 × 10⁵ cells/well. Leukotrienes in cell culture supernatants were determined as described above.

Macrophages (1 × 10⁶ cell/ml) in RPMI 1640 containing 10% heat-inactivated FCS were seeded at 100 μl/well in a 96-well black plate, with a clear bottom (no. 3603; Corning) and incubated for 1 h at 37°C. One hundred microliters of fluorescent-opsonized Zymosan from Saccharomyces cerevisiae (Molecular Probes) was added to reach a ratio of 100 bioparticles/cell. The plate was incubated at 37°C for various time points. Then the suspension was removed by aspiration, 100 μl of trypan blue was added per well, allowed to incubate for 1 min at room temperature, and then removed by aspiration. The plate was finally read at Ex 480 nm/Em 520 nm (cutoff 515 nm) on a Sprectramax Gemini device.

This was performed using the VEGF immunoassay system (BioSource International) according to the manufacturer’s instructions.

Bone marrow cells were recovered from excised murine femurs. Bones were washed three times in PBS. Femur ends were cut and bone marrow cells were flushed out with 2 ml of RPMI 1640 (10% heat-inactivated FCS) per femur using a 3-ml syringe (BD Biosciences 300699) and a microlance (25GI from BD Medical Systems). Cells were collected in a 50-ml conical tube and passaged through a cell streamer (70 μm) (BD Falcon). Afterward, cells were counted and seeded into 10- cm plastic dishes in 10% FCS-containing RPMI 1640 at a density adjusted to 10⁷ cells/ml. The next day nonadherent cells were recovered and counted again. Cells were seeded (1 × 10⁴ cells/well of a 96-well plate) and incubated in medium supplemented with 40 ng/ml recombinant murine M-CSF (R&D Systems). After 3 days, medium and nonadherent cells were discarded and cells were washed once with PBS. Fresh medium, supplemented with M-CSF (40 ng/ml), was added and cells were incubated for another 4 days. Experiments were initiated on day 7. For viability assays, we used a BrdU-based cell proliferation ELISA following the supplier’s manual (Roche). Cells were incubated with BrdU-containing medium for 2 h at 37°C/5% CO₂ in a humidified atmosphere. An anti-BrdU Ab was incubated with the samples for 1 h at room temperature. Samples were finally analyzed with a luminescence counter (Wallac Jet 1450 MicroBeta).

One microgram of mouse TNF-α (Stathmann Biotech) dissolved in saline was injected i.v. into the mouse tail vein. Peripheral blood was withdrawn through puncture of the retro-orbital plexus under anesthesia 30 min after TNF-α injection. Neutrophils were quantified by staining with Gr-1 (clone RB6-8C5; BD Pharmingen) and counting in parallel with calibration standards (Caltag Laboratories) using a FACSCalibur instrument (BD Pharmingen).

Animals were immunized i.p. on days 0, 7, and 16 by using 10 μg of purified OVA (Sigma-Aldrich) dissolved in aluminum hydroxide (1.3% in water, Alu Gel S; Serva) and mixed with a 0.9% saline solution in a total volume of 0.2 ml. Five days after the last immunization, anaphylaxis was elicited by injection of 30 μg of OVA into one rear footpad, while the opposite rear footpad received 25 μl of vehicle alone. Footpad swelling was measured after a time frame of 25 min.

Cerk^−/− and WT female mice were injected intradermally with 50 ng of mouse monoclonal anti-DNP IgE on both flanks 24 h before i.v. challenge with 200 μg of DNP-albumin conjugate, mixed with 0.5% Evans blue (20). Anaphylaxis-mediated extravasation was assessed 30 min after elicitation. Evans blue was extracted from biopsies from the test sites with formamide and measured photometrically.

WT and Cerk^−/− male mice were sensitized intradermally on the back at two sites to methylated BSA (mBSA; Fluka Chemie) homogenized 1:1 with CFA on days −21 and −14 (0.1 ml containing 1 mg/ml mBSA). On day 0, the right knee received 10 μl of 10 mg/ml mBSA in 5% glucose solution (Ag-injected knee), while the left knee received 10 μl of 5% glucose solution alone (vehicle-injected knee). The diameters of the left and right knees were then measured using calipers immediately after the intra-articular injections and again on days 2, 4, and 7. In drug-treated animals, vehicle (saline) at 5 ml/kg or efipladib at 30 mg/kg was administered daily by oral gavage. Right knee swelling was calculated as a ratio of left knee swelling, and the obtained swelling ratio was plotted against time for WT vs Cerk^−/− animals and for vehicle vs efipladib treatment groups.

Naive female WT and Cerk^−/− mice were anesthetized by s.c. injection of a sterile, filtrated mixture of 0.2 ml of anesthetic (ketamine, 1.03 mg/mouse, plus xylazine, 0.21 mg/mouse). Subsequently, serum from K/BxN mice, a strain that spontaneously develops arthritis, was injected i.p. to these recipient mice (250 μl/mouse). Paw swelling was assessed every 2 days on an arbitrary scale of 0–24/mouse. The tarsal and metatarsal phalanges of each paw were evaluated, each with a score of 0–3, thus obtaining maximally a score of 6 per paw. The paw scores were summed up to obtain a score for each individual animal. The individual sum scores of all of he animals were averaged. The scoring system used was: 0 = no detectable sign of inflammation; 1.0 = entire paw swollen; 2.0 = swollen paw involving wrist or ankle; and 3.0 = ankylosis or severely swollen paw. Treatment with efipladib was performed as described above.

Streptococcus pneumoniae ATCC 6303 was grown in Todd-Hewitt broth (Difco/BD Biosciences) supplemented with 30% heat-inactivated horse serum (Invitrogen Life Technologies) for 5 h at 37°C/ 5% CO₂ and adjusted to a concentration of 1.63 × 10⁶ CFU/ml by dilution in sterile 0.86% saline. A mouse pneumonia model adapted from that of Tateda et al. (21) was used. Thirty WT and Cerk^−/− mice were anesthetized by i.m. injection of a mixture of ketamine (90 mg/kg body weight; 10% Ketamidor) and xylazine (10 mg/kg; 2% Rompun) and infected intranasally by instillation of 50 μl of the bacterial suspension (8.15 × 10⁴ CFU/animal). Weight loss, bacterial counts in the lungs, and survival of the animals were followed over a period of 4 days. Animals were sacrificed 24 and 48 h (n = 7 for each group), 72 h (n = 8 for each group), and 96 h (N = all surviving animals) after infection. The lungs were removed aseptically and homogenized (Omni TH Tissue Homogenizer; Omni International) in 2 ml of saline. A volume of 100 μl of serial 10-fold dilutions of the homogenates was plated on blood agar to obtain counts of viable bacteria.

Cerk knockout (KO) in BALB/c embryonic stem cells was achieved by the substitution of Cerk exons 2–5 with a neomycin resistance gene cassette (Fig. 1,A). Cerk^−/− offsprings were validated by Southern blotting (Fig. 1,B) and PCR (Fig. 1,C). Analysis at the mRNA levels by real-time PCR was done on cerebellum (Fig. 1,D). It confirmed the absence of expression of CerK in Cerk^−/− animals and showed that the gene expression levels of other sphingolipid-metabolizing enzymes, in particular of the closely related CerK-like (22, 2) and the multisubstrate lipid kinase (3), were not modified as a result from CerK ablation. Using [³²P]ATP and an established CerK assay procedure (23) revealed no residual activity in tissues such as cerebellum or testis from Cerk^−/− mice (Table I), tissues which harbor high levels of CerK activity in WT mice (4). In tissues with lower CerK activity, such as pancreas and kidney, an 80% reduction was found in Cerk^−/− animals. The significance of the higher residual levels in these tissues (∼20%) is, however, difficult to ascertain given the detection limits of our assay. CerK activity was also determined in various cell types using NBD-C6-ceramide, a fluorescent short-chain ceramide species that accumulates at the Golgi complex (24) where various sphingolipid-metabolizing enzymes, including CerK, are present. Whether analyzed in mast cells, macrophages, or neutrophils, the results showed that there is no residual CerK activity in Cerk^−/− animals (Fig. 1,E). [³³P]orthophosphate labeling of WT and Cerk^−/− macrophages indicated that endogenous C1P species labeled under these conditions were not formed in Cerk^−/− cells (Fig. 1,F). Furthermore, we analyzed the levels of C16-C1P, a reportedly abundant C1P subspecies (25), in tissues and cells derived from WT and Cerk^−/− animals by LC followed by MS (LC/MS). Macrophages contained the highest amount of C16-C1P (45 ± 17 pmol/mg protein) we could measure among various cell types of murine origin (Table II). In cerebellum, mast cells, and macrophages, C16-C1P levels were reduced by >70% in Cerk^−/− compared with WT samples, whereas in neutrophils, a nonsignificant 20% reduction was measured (Table II). This indicates that, at least in some cell types or tissues, CerK plays a major role in the production of C16-C1P. However, because C16-C1P levels were not completely abolished, this also implies that alternative pathways to generate C1P exist.

As far as phenotype of Cerk^−/− mice is concerned, there was no obvious difference compared with WT littermates. Cerk^−/− mice were viable, fertile, and without histological abnormalities in the major organs (data not shown). Only a 10–15% reduction in body weight was noticed in adult Cerk^−/− female mice.

Macrophages can phagocytose microorganisms and have a major role in removal of cellular debris and induction of tissue repair (reviewed in Ref. 26). Phagocytosis, in particular, was reported to depend on CerK/C1P (7). Therefore, we investigated the phagocytic potential of macrophages isolated from Cerk^−/− animals toward fluorescent zymosan particles. Phagocytosis occurred at the same rate and to the same extent as observed with WT macrophages (Fig. 2,A). In response to appropriate signals, macrophages also have the capability to switch from a proinflammatory phenotype toward an anti-inflammatory one that is important for healing and remodeling of injured tissues (reviewed in Ref. 27). This second step involves angiogenesis and the secretion of vascular endothelial growth factor from macrophages. We analyzed this “angiogenic switch” (28) upon cotreatment with LPS and 5′-N-ethylcarboxamidoadenosine, an adenosine-1 receptor agonist (29).WT and Cerk^−/− macrophages responded similarly in this assay (Fig. 2,B). In fact, LPS treatment itself, independently of 5′-N-ethylcarboxamidoadenosine, led to a profound decrease in CerK activity in WT macrophages (see Discussion). This is shown here by adding NBD-C6-ceramide to the cells at the end of treatment (Fig. 2 B, inset).

Survival of BMDM was reported to rely on the exogenous addition of C1P in the absence of M-CSF (30). We therefore compared the survival rates of WT and Cerk^−/− macrophages in the presence or absence of M-CSF. M-CSF had a similar impact on WT and Cerk^−/− macrophages by expanding their life span (Fig. 2,C). There was no significant reduction of survival in Cerk^−/− BMDM when compared with WT cells, both in the presence or absence of M-CSF (Fig. 2 C).

Macrophage function was also investigated in vivo. Migration of Cerk^−/− macrophages during an induced peritonitis model was not modified compared with WT macrophages and neither was the pattern nor the extent of cytokine release after i.v. LPS injection (data not shown). In a model for idiopathic thrombocytopenic purpura, a disorder in which there is accelerated removal of platelets from the circulation, primarily mediated by splenic macrophages, no difference between WT and Cerk^−/− animals was observed (data not shown).

We then directly addressed cPLA₂α-dependent pathways because C1P produced by CerK was shown to be capable of translocation and activation of this cPLA₂ isoform and was proposed to be a missing link between cytokine signaling at the cell surface and perinuclear eicosanoid biosynthesis (11). We selected readouts, previously shown to critically depend on cPLA₂α (14, 15), questioning whether they would be affected by CerK deficiency. PMA/ionomycin treatment of peritoneal macrophages resulted in a 50% increase in arachidonic acid release (Fig. 3,A), in line with reported data (15). This response was identical in WT and Cerk^−/− mice-derived macrophages and was fully dependent on cPLA₂α activity because 50 nM efipladib, or Wyeth-1, a selective inhibitor of cPLA₂α when used at this concentration in cell-based assays (31, 32), abrogated the signal (Fig. 3,A). In line with this observation, the extent of PGE₂ release, which is positively correlated with arachidonic acid production, was similar in WT and Cerk^−/− macrophages, and efipladib completely abrogated the induced release (Fig. 3,A, right). LPS was previously shown to depend, for a large part, on cPLA₂α to increase PGE₂ production. This was confirmed here and the cPLA₂α-dependent component of the signal again appeared to be comparable in WT and Cerk^−/− macrophages, as seen with the extent of inhibition with efipladib (Fig. 3,B). Similar findings were obtained with primary kidney fibroblasts (data not shown). IL-1β was previously reported to stimulate PGE₂ production as a result of C1P production by CerK in A549 cells (33); therefore, we tested this cytokine in peritoneal macrophages. IL-1β treatment led to a modest increase in PGE₂ production, which appeared similar in WT and Cerk^−/− cells (data not shown). Finally, we evaluated the release of leukotrienes in response to treatment with the calcium ionophore A23187, which is considered to be fully dependent upon cPLA₂α (14, 15). Again, Cerk^−/− cells responded similarly to WT cells (Fig. 3 C).

The possibility of a link between CerK and cPLA₂α-dependent pathways was further checked in vivo using models for rheumatoid arthritis, a disease known to critically depend on cPLA₂α (34). Models of Ag-induced arthritis (12) (Fig. 4,A) and of arthritis induced by serum transfer (13) (Fig. 4,B) were performed to compare the responses of Cerk^−/− mice to those of WT animals (Fig. 4, left) or to compare treatment with efipladib vs placebo in WT mice (Fig. 4, right). Whereas efipladib strongly and significantly reduced joint swelling in the diseased animals, Cerk^−/− disease scores were not significantly lower than those of WT animals in both models. Therefore, these in vivo experiments further demonstrate that cPLA₂α pathways are functional in Cerk^−/− animals.

CerK was recently reported to play a role in mast cell function (8, 9). We therefore investigated in vitro the three phases of mast cell signaling following IgE/Ag stimulation: degranulation, triggering of the arachidonic acid pathway, as well as cytokine release (Fig. 5,A and data not shown). None of these readouts were modified in cells from Cerk^−/− animals, as compared with WT cells. In vivo models known to depend on mastocyte function, such as active or passive cutaneous anaphylaxia, were also undertaken. These represent disease models for systemic or local immediate hypersensitivity reactions, respectively. There was no difference between Cerk^−/− and WT mice in their susceptibility to active anaphylaxia (Fig. 5,B). The response to passive anaphylaxia, however, was lowered by 32% in Cerk^−/− animals (Fig. 5 C).

The possible role of CerK and/or C1P in neutrophil functions was pointed out in previous studies (7, 35, 36). We analyzed leukocytes in bone marrow, spleen, and blood by differential staining and FACS analysis. In bone marrow, the proportion of granulocytes was similar for WT and Cerk^−/− animals (Fig. 6,A and Table III). However, in spleen and blood, their proportion was decreased by at least 50% in Cerk^−/− animals (Fig. 6,A and Table III). In bone marrow and spleen, the amount of leukocytes other than neutrophils appeared to be increased in Cerk^−/− animals; this did not translate into significant organomegaly (data not shown). To better understand the neutropenia pattern of Cerk^−/− animals, we performed repeated blood cell measurements over a period of 4 wk. This study did not reveal any sign for cyclic neutropenia in Cerk^−/− animals. Furthermore, a slight increase of lymphocytes always compensated for the reduction of neutrophils in Cerk^−/− animals, confirming the data shown in Fig. 6 A, whereas monocytes counts were identical to those of WT animals (data not shown).

Despite their lower abundance, the capacity of Cerk^−/− neutrophils to leave the bloodstream and adhere to the endothelium following its activation by TNF-α was not compromised (Fig. 6,B). Cerk^−/− neutrophils were also able to be produced and migrate to sites of inflammation, as seen in an induced peritonitis model (Fig. 6 C), implying that the acute generation of neutrophils and their migratory capacity was preserved in the absence of CerK.

Neutrophils have a short life span and therefore we wondered whether neutropenia might be occurring in Cerk^−/− animals as a result of increased cell death. Mechanisms that regulate neutrophil apoptosis are still poorly understood but there has been increasing awareness that ceramide levels may play an important regulatory role as inducer of neutrophil apoptosis (36, 37). We, therefore, measured ceramide species across a broad range of chain lengths in sera from WT and Cerk^−/− animals. Not only were ceramide levels significantly increased in Cerk^−/− compared with WT animals but the corresponding dihydroceramide species were strongly reduced (Table IV). Consequently, the ratio of ceramide:dihydroceramide was increased seven times overall in Cerk^−/− compared with WT animals. Such an increase of ceramide at the expense of dihydroceramide indicates that Cerk^−/− animals may be primed for apoptosis and this defines an important cause underlying neutropenia.

Since the clinical result of neutropenia is increased risk for infection, we submitted WT and Cerk^−/− mice to a model of fulminant pneumonia. Until day 2 postinfection, no difference was noticeable between WT and Cerk^−/− animals. However, from day 3 onward, the course of disease worsened in Cerk^−/− animals. Rough coat, increased breathing rate, and decreased activity were noticed in Cerk^−/− animals, and we measured more significant body weight loss (Fig. 7,A) as well as a larger rise in lung bacterial counts in these animals (Fig. 7,B). Finally, at day 4, the survival rate of Cerk^−/− animals had dropped to 25%, whereas all WT were still alive (Fig. 7 C).

In this work, we have addressed the consequences of CerK ablation using a comprehensive experimental framework aimed at probing innate immune cell function. We first performed an extensive analysis to check for CerK activity and C1P levels in Cerk^−/− animals using all currently available assays. This unequivocally showed that Cerk^−/− animals are devoid of CerK activity (Fig. 1, E and F, and Table I). C16-C1P levels are significantly reduced in these animals, but not fully abolished, pointing to another possible, maybe compensatory, source of C1P (Table II). This study, therefore, provides the first evidence for the possible existence of different pools of C1P resulting from alternative synthesis pathway(s). Therefore, the strain described here will be a valuable tool to search for additional C1P-generating activities. Whether and to what extent compensation might be also occurring upon acute CerK inhibition will also be important to explore once a potent and selective CerK inhibitor has been identified.

Not only C1P levels were influenced by ablation of CerK. Ceramide levels in the serum of Cerk^−/− animals were up-regulated (Table IV). Concomitantly, levels of dihydroceramides (the precursors of ceramides in biosynthesis which are devoid of proapoptotic activity) were reduced, maybe as a feedback response to the increased ceramide levels (Table IV). Sphingosine and sphingosine-1-phosphate levels were unchanged in the serum of Cerk^−/− animals compared with WT animals (data not shown). Therefore, this study has revealed that CerK plays a role in the regulation of ceramide levels with impact on the conversion of dihydroceramide into ceramide.

We analyzed macrophage function in Cerk^−/− animals using an extensive panel of phenomenological readouts. First, we analyzed in vitro the phagocytic capacity and the ability to secrete vascular endothelial growth factor, which represent pro- and anti-inflammatory responses in macrophages, respectively, and did not detect differences in the absence of CerK (Fig. 2). Interestingly, LPS treatment of WT macrophages led to a profound decrease in CerK activity (Fig. 2 B, inset). To our knowledge, this is the first evidence to date for such a profound down-regulation of CerK activity, and we are in the process of characterizing this further. Then, we addressed in vivo 1) the ability of macrophages to migrate into the peritoneal cavity during peritonitis; 2) the profile of cytokine secretion in blood following i.p. injection of LPS, which represents a functional readout for macrophages; and 3) performed a disease model of idiopathic thrombocytopenic purpura: in all models, responses in Cerk^−/− mice were identical to those in WT animals (data not shown).

CerK and C1P were previously shown to be important for cell proliferation and survival (reviewed in Ref. 38). It is clear that ablation of CerK has not prevented Cerk^−/− animals to grow normally, to reach adult size and perform, in all our assays, in an almost undistinguishable way as compared with WT littermates. However, the slight reduction in body weight observed in Cerk^−/− animals speaks for a contribution of CerK to growth overall. Murine BMDM display high CerK activity (Fig. 1,E). Using BMDM from CD-1 mice, Gomez-Munoz et al. (30) have reported the antiapoptotic effect of exogenously added C1P upon M-CSF withdrawal. We observed a shorter survival time when WT BMDM were deprived of M-CSF, consistent with literature data (38, 39). However, no further shortening of the life span in BMDM was observed in the absence of CerK (Fig. 2 C). Therefore, residual levels of C1P in Cerk^−/− BMDM, possibly obtained through compensatory mechanisms (see above), may have been sufficient to prevent accelerated cell death. An alternative possibility may be the sensitivity to apoptosis of BMDM from BALB/c mice, which may not be as pronounced as that of cells from CD-1 mice because the latter are unable to produce M-CSF. Further studies will be required to precise the role of CerK and C1P in cell growth.

In macrophages, cPLA₂α is instrumental to the generation of eicosanoids, upon various stimuli (14, 15). Macrophages, therefore, represent a cell type of choice to investigate the ability of CerK/C1P to activate this pathway. Although significantly reduced levels of C1P were found in macrophages isolated from Cerk^−/− animals (Fig. 1,F and Table II), cPLA₂α pathways were undisturbed (Fig. 3). In case C1P is indeed able to stimulate cPLA₂α in these cells, this would imply that residual or compensatory C1P levels may be sufficient and able to localize to sites where cPLA₂α and COX-2 enzymes are functional, or that C1P is not essential and can be efficiently substituted by another lipid mediator. Moreover, disease progression in rheumatoid arthritis models, which depends upon cPLA₂α (34), was not significantly impaired in Cerk^−/− mice as compared with WT littermates, as seen using two different experimental settings (Fig. 4). Altogether, cPLA₂α -dependent pathways function normally in murine Cerk^−/− macrophages, therefore not withstanding the proposal for a key role of CerK-derived C1P in the regulation of cPLA₂α (11).

A previous study linked CerK/C1P to the process of mast cell degranulation using CerK-overexpressing RBL cells (8). We have looked at a number of parameters covering all phases of mast cell activation in BMMC. We did not detect any difference between cells from WT and Cerk^−/− animals, despite significantly reduced C1P levels in the latter (Table II and Fig. 5,A). In vivo, models known to critically rely on mast cell function and used as models for allergic diseases also did not show any (active cutaneous anaphylaxia, Fig. 5,B) or striking (passive cutaneous anaphylaxia, Fig. 5 C) differences between Cerk^−/− and WT animals. Therefore, a role of CerK in mast cell biology remains to be demonstrated.

A clear phenotype of Cerk^−/− animals is the reduced number of neutrophils in the circulation and to some extent also in the spleen (Fig. 6). Altogether, our data indicate that neutropenia is, at least in part, due to an increased rate of neutrophil apoptosis in CerK-deficient animals: 1) ceramide, either produced within the cells (37) or added to the medium (40) is known to induce apoptosis and actually appears to be a crucial regulator of programmed cell death in this cell type and 2) we show here that ceramide levels in the serum (and, although not measured, probably also within cells) are up-regulated in Cerk^−/− animals (Table IV). Neutropenia does not result from sequestration of neutrophils or decreased bone marrow production since neutrophils can be made and recruited in numbers in Cerk^−/− animals as seen in a model of peritonitis (Fig. 6 C). However, we cannot exclude the possibility of diminished capacity of Cerk^−/− neutrophils to exit the bone marrow under normal conditions.

Neutropenia is associated with higher risk for infection; therefore, we examined the capacity of WT and Cerk^−/− animals to fight an infection challenge with S. pneumoniae, where neutrophils play a critical role. After a lag phase, the disease course was strikingly accelerated in Cerk^−/−, resulting in significant changes in experimental scores as well as a rise in mortality at the end of the experiment (Fig. 7). Therefore, despite current lack of compelling in vitro evidence for a role of CerK in neutrophil function (36), the neutropenia found in Cerk^−/− mice as well as their impaired defense against pulmonary infection supports further investigation in this cell type.

In conclusion, we report here on the generation of Cerk^−/− mice, on their characterization in terms of CerK activity, C1P and related metabolite levels, as well as on the results of an extensive profiling to probe their innate immune responses. This work has come with important findings: 1) the large contribution of CerK to C1P levels, in particular with a C16 side chain length; 2) the presence of residual C16-C1P in the absence of CerK, an indication that other source(s) of C1P likely exist; 3) the strong increase in ceramide species at the expense of dihydroceramide that establishes a proapoptotic milieu when CerK is absent; 4) the finding that CerK is dispensable for innate inflammatory responses in macrophages and mast cells, in particular, responses where cPLA₂α is known to be instrumental, but that CerK is important for neutrophil homeostasis; and 5) the impaired defense against S. pneumoniae infection in the absence of CerK. Whether CerK may serve as a valuable therapeutic target to treat inflammatory disorders, as suggested previously (11), is not obvious at this stage. This is reminiscent of the situation regarding sphingosine kinases where data obtained in KO animals also challenged a number of previously established hypotheses (41, 42). The further profiling of Cerk^−/− mice, in particular to identify other sources for C1P formation, will remain the subject of continuing study.

We thank the following Novartis Institutes for Biomedical Research colleagues who helped with the experimental part: Alistair Boath for isolation of organs, Julie Boisclair and her laboratory for the histopathological analyses, Robert Csonga for experiments with mast cells, Eva-Marie Haupt for ceramide measurements, Werner Hoellriegl and his team for animal care taking, Raphaela Kutil for multiplex cytokine measurements, Edith Pursch for the preparation of mast cells, and Alexander Wlachos for RT-PCR measurements. We also thank Marijke Nefzger for help with statistic evaluations. We are grateful to A. Bielawska (Medical University of South Carolina, Charleston, SC) for providing C16-C1P.

The authors have no financial conflict of interest.