Research Paper Mediators of Inflammation, 8, 31–35 (1999)

ARACHIDO N IC acid (AA), a fatty acid found in the human bone marrow plasma, is the precursor of eicosanoids that modulate bone marrow haematopoiesis. To further our understanding of the role of AA in the bone marrow physiology, we have assessed its incorporation in human bone marrow mononuclear cells. Gas chromatography analysis indicates the presence of AA in their fatty acid composition. In bone marrow mononuclear cells, [ 3 H]-AA is incorporated into tri-glycerides and is later delivered into phospholipids, a result not observed with blood mononuclear cells. Prelabelling-chase experiments indicate a trafficking of labelled AA from phosphatidylcholine to phospha-tidylethanolamine. Stimulation of prelabelled bone marrow mononuclear cells with granulocyte-macro-phage colony-stimulating factor (GM-CSF) results in the release of a part of the incorporated labelled AA. Finally, exogenous AA (up to 1 m M) has no significant effect on cell growth. In conclusion, human bone marrow mononuclear cells participate to the control of marrow AA concentrations by incorporating AA into phospholipids and triglycerides. In turn, bone marrow mononuclear cells can release AA in response to the potent haematopoietic growth factor GM-CSF.


Introduction
Arachidonic acid (AA), a 20-carbon polyunsaturated fatty acid, is converted to prostaglandins through the cyclooxygenase pathway or to hydroxyeicosatetraenoic acid (HETE) and to leukotrienes through the lipoxygenase pathway. 1 Several AA metabolites modulate human haematopoiesis by acting on the growth of human bone marrow erythroid and myeloid progenitors, 2 -5 and on the cytokine production by human bone marrow stromal cells. 6 The role of AA in the bone marrow physiology is poorly documented. Recently we have reported the presence of AA in the human bone marrow plasma and its capture by cultured human bone marrow stromal cells. 7 To further our understanding of the role of AA in the bone marrow, the uptake and the distribution of AA among triglycerides (TG) and phospholipids (PL) in human bone marrow mononuclear cells were compared with those in blood mononuclear cells and erythrocytes. We have also investigated the role of AA on the proliferation of human bone marrow mononuclear cells and their capacity to release AA in response to granulocyte-macrophage colony stimulating factor (GM-CSF), a potent haematopoietic growth factor.
Human bone marrow cells from patients undergoing a myelogram as part of a diagnostic procedure were harvested by aspiration into heparinized tubes (Vacutainer system, Becton Dickinson, Meylan, France). When samples were considered as normal by morphological analysis of the May Grunwald Giemsa stained smears, bone marrow mononuclear cells were isolated by separation on a Ficoll gradient (400 3 g, 20 min) and washed twice with HBSS (400 3 g, 10 min) before use. For preparation of blood mononuclear cells, human blood was drawn from healthy volunteers using heparin as anticoagulant and prepared as bone marrow cells. For preparation of erythrocytes, blood was recovered using heparin as anticoagulant. The buffy coat was aspirated and discarded. The red cells were washed three times with HBSS before use.

Fatty acid analysis of human bone marrow mononuclear cells
Cell lipids were extracted from freshly isolated mononuclear bone marrow cells according to Floch et al. 8 Fatty acids were then methyl transesterified and the resulting fatty acid methyl esters were separated by gas chromatography. 9 Standards were run to identify the compounds quantified by gas chromatography.

Lipid extraction
Cells were mixed with 1 ml of 0.1% sodium dodecyl sulphate in water, and incubated for 30 min at 56°C with 3 ml of chloroform/methanol (2:1, v/v). The chloroformic extract was washed with 5 ml of 0.1 M KCl/methanol/chloroform (96:92:6, v/v/v), and was evaporated. 10 The dry extract was recovered in 150 m l of chloroform and applied to a thin layer chromatography (TLC) plate (Silica gel 60 (20 3 20 cm, 0.25 mm), Merck) and submitted to TLC. The plate was developed in the mixture of diethyl ether/ hexane/acetic acid (70:30:1, v/v/v). Each lane was divided in areas of 0.5 cm length which were scraped into vials and radioactivity was measured on a Packard liquid scintillation counter. Solutions of MG, DG, TG, PL and AA were used as standards and visualized with iodine vapour.

Analysis of labelled PL
For the separation of the various species of cellular PL, the corresponding areas were scraped from the TLC plates and were extracted with chloroform/ methanol (2:1, v/v). Samples were then rechromatographed using a solvent system of chloroform/ methanol/acetic acid/water (50:30:8: 10 Each lane was divided in areas of 0.5 cm length and processed as above. Solutions of PC, PS, PI, and PE were used as standards and visualized with iodine vapour. In this chromatography, PI and PS migrated to the same area. To ensure that the labelled compound in cellular PL from freshly-isolated marrow mononuclear cells was authentic AA, labelled PL were submitted to base hydrolysis. 11 The labelled PL were resuspended in 1 ml of 2 M KOH in ethanol/water (3:1, v/v). After 40 min at 60°C, 1 ml of water and 1 ml of 6 N HCl were added to the mixture to acidify the phase (pH 3). Labelled compounds were extracted with chloroform/methanol (2:1, v/v) and rechromatographed using a solvent system of diethyl ether/hexane/acetic acid (70:30:1, v/v/v). The amount of radioactivity migrating with PL and free AA was determined by liquid scintillation counting. In another set of experiments labelled PL were hydrolysed with PLA 2 from bovine pancreas (Sigma). Briefly, labelled PL samples were resuspended in 1 ml of diethyl ether/methanol (95:5, v/v) containing 50 m l of Tris buffer (0.05 M), CaCl 2 (25 mM), EDTA (1 mM), and 1 mg of phospholipase. The reaction mixture was incubated at 37°C for 2 h with continuous shaking. After evaporation of solvent, samples were rechromatographed as for base hydrolysis.

Proliferation studies
Freshly-isolated bone marrow mononuclear cells (1 3 10 5 per well) were cultured in 96-well microtitre plates in 100 m l of RPMI 1640 with 10% FCS in the presence of AA (from 1 m M to 0.1 nM) or the appropriate vehicle (10 m l 2% HSA). After 64 h of incubation, cultures were pulsed for 8 h with 1 m Ci/ ml [ 3 H]thymidine and cells were harvested using a Skatron cell harvester. Each experiment was carried out as sixplicates. In separate sets of experiments, cells (in triplicate samples) were harvested and counted by using a haemocytometer. Data were compared using the Student's t-test for paired variables.

Release of AA from [ 3 H]-labelled bone marrow mononuclear cells
Freshly-isolated human bone marrow mononuclear cells were labelled for 0.5 h with [ 3 H]-AA as reported above. After washing, labelled cells (1 3 10 6 ) were suspended in 1 ml of RPMI 1640 containing 0.25% fatty acid-free BSA, 1 mM Ca 2+ and 1 mM Mg 2+ . Cells were then incubated at 37°C in the absence or presence of GM-CSF (100 ng/ml). After various periods of time (from 1 to 60 min) the mixture was centrifuged (5 min, 2000 3 g) and the [ 3 H]-supernatant fluid was removed into a separate vial. Radioactivity in the packed labelled cells and in the radioactive supernatant fluid were then measured on a Packard liquid scintillation counter.

Fatty acid composition of human bone marrow mononuclear cells
As shown in Table 1, AA is present in freshly isolated human bone marrow mononuclear cells.

Incorporation of [ 3 H]-AA in cellular lipid species
After 5 h of incubation with [ 3 H]-AA, freshly-isolated bone marrow mononuclear cells, blood mononuclear cells and erythrocytes had incorporated 52%, 49% and 41% of the initially added radioactivity, respectively (mean of three experiments). Lipids were extracted from labelled cells and separated into classes by TLC. The distribution of label in cell lipids as a function of time is shown in Fig. 1. A trafficking of labelled AA from TG to PL was observed with freshly-isolated human bone marrow mononuclear cells (Fig. 1A) but not with blood mononuclear cells and erythrocytes ( Fig. 1B and 1C, respectively). For mononuclear bone marrow cells at 24 h, 67 ± 8% of the label was recovered in PL while 30 ± 7% was recovered in TG. Less than 3% of label was recovered in DG or as free AA. Labelled MG were not detected.

Incorporation of [ 3 H]-AA in cellular PL species
We next investigated the distribution of [ 3 H]-AA into PL species using mononuclear bone marrow cells labelled for 0.5 h. The percentage (mean of three experiments) of incorporated AA into PE (43.7 ± 4.3%) was similar to that for PC (42.0 ± 3.6%) and higher than for PI + PS (14.1 ± 1.6%). We then investigated the putative exchange of labelled AA between PL species. Freshly-isolated bone marrow   (Fig. 2). Base hydrolysis of labelled PL from bone marrow mononuclear cells resulted in the loss of 99% (mean of three experiments) of radioactivity, which was recovered as free fatty acid. Similarly, treatment of labelled PL with PLA 2 from bovine pancreas lead to the loss of 80% (mean of two experiments) of label.

Release of [ 3 H]-AA by bone marrow mononuclear cells
Prelabelled human bone marrow mononuclear cells were challenged with GM-CSF (100 ng/ml) for various periods of time. As reported in Fig. 3 Effect of AA on the growth of human bone marrow mononuclear cells Table 2

Discussion
AA metabolism plays a pivotal role in many cellular functions, 1 and AA metabolites act on human marrow haematopoiesis in vitro. 2 -5 Recently we have reported the presence of AA in the human bone marrow plasma. 7 To further our understanding of the role of AA in the human bone marrow, we have investigated the capacity of human bone marrow mononuclear cells to capture and to release AA as well as the putative role of AA in their growth.
Using gas chromatography analysis, we found that AA is present in freshly-isolated human bone marrow mononuclear cells. Similarly to what was observed with human bone marrow stromal cell cultures, 7 experiments with [ 3 H]-AA indicate that freshly-isolated bone marrow mononuclear cells incorporate AA into PL and TG. Time-course experiments reveal a transfer of AA from TG to PL. As already shown in several inflammatory cell types, the incorporation of AA into TG is important to capture fatty acids for PL biosynthesis but also to regulate AA metabolism. 12,13 In contrast to bone marrow mononuclear cells, no transfer of AA from TG to PL is observed with blood mononuclear cells. As previously reported for various kinds of mammalian erythrocytes, 14,15 human erythrocytes incorporate exogenous AA into PL and TG. Their lack of transfer of AA from TG to PL might be related to their inability of de novo PL synthesis. 12 The analysis of PL species of human bone marrow mononuclear cells indicates a preferential incorporation of AA into PE and PC. Similarly to what was observed in several inflammatory cell types, 12 of labelled AA from PC to PE. At this time the meaning of this traffic is unclear. PLA 2 -treatment and base hydrolysis experiments confirm that [ 3 H]-AA is incorporated unmodified in mononuclear bone marrow PL and is linked by an ester bond. Stimulation of prelabelled bone marrow mononuclear cells with GM-CSF results in the hydrolysis of a part of the incorporated labelled AA. This effect might be mediated through the activation of a cellular phospholipase A 2 (PLA 2 ) activity. Strenghtening this hypothesis GM-CSF generates PLA 2 activity with the release of free fatty acids from neutrophil plasma membranes. 17 Exogenous AA does not change the growth of human bone marrow mononuclear cells assessed by [ 3 H]-thymidine incorporation into DNA and cell counts. In similar experimental conditions several lipoxygenase metabolites of AA have no effect on their growth. 18 As a positive control of cell stimulation we may mention that in similar experimental conditions, the lipid mediator platelet-activating factor (PAF) stimulates their [ 3 H]-thymidine incorporation. 19 In conclusion human bone marrow mononuclear cells might participate to the control of marrow AA concentrations by incorporating AA into their PL and TG. In turn, the potent haematopoietic growth factor GM-CSF stimulates their AA release, a result of importance considering the role of eicosanoids during human marrow haematopoiesis. 2