Human Eosinophils Express, Relative to Other Circulating Leukocytes, Large Amounts of Secretory 14-kD Phospholipase A₂

Formyl-methionyl-leucyl-phenylalanine (fMLP), cytochalasin B (cyto B), phorbol myristate acetate (PMA), and calcium ionophore A 23187 (Sigma Chemical Co, St Louis, MO) were dissolved in dimethyl sulfoxide (DMSO) at 1,000 times the final concentration for cell incubations. PAF (Sigma) was dissolved in phosphate-buffered saline (PBS) containing human-serum albumin (HSA) (0.5% vol/vol). Interleukin-5 (IL-5) was obtained from Amersham Corp (Amersham, Birmingham, UK). STZ was prepared as described previously.9 Serum-opsonized Sephadex particles (SOS) were prepared by incubating swollen Sephadex G15 beads (150 mg) in 1 mL of human AB serum for 30 minutes at 37°C followed by three washes in 0.9% (wt/vol) NaCl. Washed beads were resuspended in incubation medium (see below) to a concentration of 150 mg/mL. The monoclonal antibodies (MoAbs) 4A1 and 10B2 raised against sPLA₂ were kindly provided by Dr F.B. Taylor, Jr (Oklahoma Medical Research Foundation, Oklahoma City, OK). The polyclonal anti-sPLA₂ rabbit antiserum was purchased from Cayman Chemical Co (Ann Arbor, MI). All other chemicals were reagent grade. Incubation medium for the cell incubations contained 132 mmol/L NaCl, 6.0 mmol/L KCl, 1.0 mmol/L CaCl₂, 1.0 mmol/L MgSO₄, 1.2 mmol/L potassium phosphate, 20 mmol/L HEPES, 5.5 mmol/L glucose and 0.5% (wt/vol) human albumin, pH 7.4.

Blood was obtained from healthy volunteers. Granulocytes were isolated from the buffy coat of 500 mL of blood anticoagulated with 13 mmol/L trisodium citrate (pH 7.4) essentially as described before.21 In short, the mononuclear cells were removed by centrifugation of the blood over isotonic Percoll (1.077 g/mL, pH 7.4). After isotonic lysis of the erythrocytes in the pellet with an ice-cold solution containing 155 mmol/L NH₄Cl, 10 mmol/L KHCO₃, and 0.1 mmol/L EDTA (pH 7.2), the granulocytes were washed twice in PBS and suspended in incubation medium. Where indicated, neutrophils were further purified by depletion of the eosinophils with CD9 MoAbs and magnetic Dynabeads, coated with sheep-antimouse antibodies. The resulting preparations always contained less than 0.2% eosinophils.

Eosinophils were purified as previously described.21 In short, granulocytes were incubated for 30 minutes at 37°C to restore the initial density of the cells. After this incubation period, the cells were washed and resuspended in PBS supplemented with HSA (0.5% wt/vol) and trisodium citrate (13 mmol/L). After preincubation of the granulocytes for 5 minutes at 37°C, fMLP (10 nmol/L) was added to the cell suspension, and the incubation was continued for 10 minutes. Subsequently, the eosinophils were purified by centrifugation (20 minutes at 800g) over isotonic Percoll (1.082 g/mL, pH 7.4), washed and resuspended in incubation medium. This suspension contained for >98% eosinophils.

Monocytes and lymphocytes were isolated from the upper fraction of the Percoll gradient (1.077 g/mL) (the mononuclear cells) by countercurrent centrifugal elutriation as described previously by De Boer and Roos.22 The purities of both cell fractions were > 90%. Basophil purification was performed by depletion of residual contaminating cells (mainly lymphocytes and a few monocytes) from the last elutriation fraction, with antibodies against these cells (CD2, CD14, CD16, CD19) and magnetic Dynabeads as described.23The purity of the basophils in this preparation was > 92%. The mononuclear cell fractions were kept on ice to prevent activation.

Human granulocytes (containing about 5% eosinophils), and purified eosinophils, basophils, lymphocytes, and monocytes were centrifuged on microscope slides and fixed in acetone for 10 minutes. All incubation steps were performed at room temperature. The preparations were preincubated with normal goat serum (1:10 diluted in PBS, containing 1% [wt/vol] bovine serum albumin [BSA]) for 15 minutes and were subsequently incubated for 1 hour with the anti-sPLA₂ MoAb 4A1 (25 μg/mL). Control preparations were incubated with PBS/BSA without MoAb 4A1 or with irrelevant MoAb of the appropriate subclass. After each incubation step,the preparations were washed with PBS. Next, the preparations were incubated with 0.3% (vol/vol) H₂O₂, 0.1% (wt/vol) sodium azide in PBS for 20 minutes to block endogenous peroxidase activity. Subsequently, the preparations were incubated for 30 minutes with biotinylated rabbit antimouse-Ig F(ab)₂ fragments (Dakopatts, Glostrup, Denmark) 1:200 diluted in PBS containing 10% (vol/vol) normal human serum (NHS). The preparations were washed and incubated for 30 minutes with horseradish peroxidase (HRP)-labeled streptavidin-biotin(AB) Complex (Dakopatts), diluted 1:200 in PBS-NHS. HRP activity was detected by using 3-amino-9-ethylcarbazole (Sigma) in the presence of 0.03% (vol/vol) H₂O₂ yielding a red color. Preparations were fixed in paraformaldehyde (PFA; 4% vol/vol), counterstained with Mayer's hematoxylin and mounted with glycerolgelatin. The preparations were examined by light microscopy, to determine the localization and intensity of staining of sPLA₂. The specificity of this procedure was confirmed by the absence of staining of pancreas tissue preparations (described to contain PLA₂ type I¹⁴) and positive staining of Paneth cell preparations (described to contain sPLA₂,14 results not shown).

Concentrations of sPLA₂ in cell lysates or cell supernatants were determined with an ELISA modified from Smith et al.24 MoAb 10B2 and 4A1 against sPLA₂ were used as catching and detecting antibody, respectively. In some experiments, the polyclonal anti-sPLA₂ rabbit antiserum was used as catching antibody. The amount of sPLA₂ was assessed by comparison with a (sigmoidal) calibration curve obtained with purified recombinant human sPLA₂ (kindly provided by Dr C. Schalkwijk, Centre for Biomembranes and Lipid Enzymology, University of Utrecht, Utrecht, The Netherlands). The sensitivity of the measurements in this assay amounted to 0.2 to 5 ng/mL. Lysis of the different cell types (2 × 10⁷ cells/mL) was performed for 30 minutes at 4°C in incubation medium with the following additions: 1% NP-40 (vol/vol), phenylmethyl sulfonyl fluoride (PMSF; 100 μg/mL), tosyl-leucyl chloromethylketone (TLCK; 1 mmol/L), leupeptin (0.1 mmol/L), soybean trypsin inhibitor (SBTI; 20 μg/mL) (Sigma), and EDTA (5 mmol/L).

Human eosinophils or neutrophils (depleted for eosinophils) were resuspended (10⁸ cells/mL) in PBS containing 0.34 mol/L sucrose, 10 mmol/L HEPES, 100 μg/mL PMSF, 1 mmol/L TLCK, 0.1 mmol/L leupeptin, and 5 mmol/L EDTA (final pH 7.0), sonicated at 21 kHz and 8 μm peak-to-peak amplitude for 3 × 15-second intervals at 4°C and stored at −80°C until use. sPLA₂ELISA measurements of these sonicates show identical amounts of sPLA₂ protein compared with lysis in 1% NP-40 as described above (data not shown). sPLA₂ activity measurement was performed with a commercially available assay system (Cayman Chemical Company) exactly as recommended by the manufacturer. The assay uses the 1,2-dithio analog of diheptanoyl phosphatidylcholine, which serves as a substrate for most PLA₂ with the exception of group IV cytosolic PLA₂. On hydrolysis of the thio ester bound at the sn-2 position by PLA₂, free thiols are detected by reaction with 5,5′-dithio-2-nitrobenzoate (DTNB). The hydrolysis of DTNB was quantified by measuring the absorbance at 405 nm at different time points and determination of the slope, using 10.0 mmol⁻¹ · L⁺¹ · cm⁻¹ as the adjusted (for pathlength) molar extinction coefficient.

ECP in cell supernatants was measured with a slightly modified sandwich ELISA previously described by Reimert et al.25 Specific polyclonal rabbit-antihuman ECP antibodies were obtained by immunization of rabbits with highly purified ECP.26 Plates were coated overnight (o/n) with rabbit-antihuman ECP (3 μg IgG/mL). After blocking with BSA (2%, wt/vol), samples and standard (purified ECP) were incubated in various dilutions for 90 minutes at 37°C in a buffer (pH 7.5) containing 0.1 mol/L NaCl, 50 mmol/L NaH₂PO₄, 0.1% Tween-20, 0.1% CTAB, 0.2% BSA and 20 mmol/L EDTA. Bound ECP was detected after incubation with biotin-conjugated rabbit-antihuman ECP by adding a mixture of avidin and biotinylated alkaline phosphatase according to the manufacturer's description (ABC-Complex Dakopatts). Enzymatic activity was determined with phosphatase substrate tablets (Sigma 104, Sigma). The range of the measurements was 0.10 to 10 ng/mL.

Eosinophils were prewarmed for 10 minutes, primed with IL-5 (10^-10 mol/L) for 30 minutes, and incubated with STZ for 10 minutes. All cells were fixed for 2 hours at room temperature in a graded (2% to 8%) PFA series in PBS to preserve the antigenicity of sPLA₂ and were pelleted in 10% gelatin in PBS. This fixation procedure was found to be optimal for staining with antibodies directed against sPLA₂ (see below). The gelatin slab was kept at 4°C in PBS containing 1% PFA. Small pieces of gelatin were infiltrated with 2.3 mol/L sucrose in PBS for 4 hours, mounted on copper pins at 4°C, and frozen in liquid nitrogen. About 80 nm cryopreparations were made on an MT-7 ultracryomicrotome (Research and Manufacturing Co, Tucson, AZ) and incubated at room temperature with mouse MoAb anti-sPLA₂ 10B2 (dilution 1/50) followed by incubation with rabbit antimouse (RAM)-IgG and goat-antirabbit (GAR)-IgG conjugated to 10 nm gold (both at dilution 1/40) (Amersham). As a control, anti-sPLA₂ was replaced by a nonrelevant murine IgG. All incubations lasted 1 hour. After immunolabeling, the cryopreparations were embedded in a mixture of methylcellulose and uranyl acetate. All preparations were examined with a Philips CM10 electron microscope (Eindhoven, The Netherlands).

The MoAb 4A1 raised against recombinant human sPLA₂ was used to determine the presence of the 14-kD sPLA₂ in circulating leukocytes. As positive control for the histochemical staining procedure, HEPG2 cells, which are known to express sPLA₂,27 were included in the experiments. The few cells staining positive for sPLA₂ in the granulocyte fraction (Fig 1A) had the appearance of eosinophils. Indeed, in cytospins of purified eosinophils, all cells stained positive for sPLA₂ (Fig 1B). In Fig 1B, it can be seen that the sPLA₂ present in the eosinophils is granular distributed. The same pattern of staining was seen in the HEPG2 cells (data not shown). The immunocytochemical preparations of neutrophils, monocytes, lymphocytes, and basophils did not exhibit a detectable staining (data not shown).

To quantitate the level of sPLA₂ in eosinophils, a sandwich ELISA with MoAbs directed against sPLA₂ was applied. Previous studies have shown that this assay is specific for the human sPLA₂ enzyme24 and that a strong correlation exists between PLA₂ activities and immunoreactive material in plasma samples of patients undergoing IL-2 therapy.28 Table 1 shows that the content of sPLA₂, present in the lysates of eosinophils was 20-fold to 100-fold higher than that of the other circulating leukocytes. Although sPLA₂ was not visualized in preparations of monocytes and neutrophils immunocytochemically stained for sPLA₂, lysates of monocytes and neutrophils (depleted for eosinophils) did contain a relative small amount of sPLA₂ (Table 1), whereas the results obtained with lysates of purified basophils were on the borderline of the sensitivity of our assay. The low signal in the neutrophil lysates could not be explained by the presence of contaminating eosinophils (being less than 0.2%). In the lysates of lymphocytes, sPLA₂ was not detected.

The pronounced difference between neutrophils and eosinophils was confirmed with ELISA measurements in which the monoclonal 10B2 antibody was replaced as catching antibody by a polyclonal anti-sPLA₂ rabbit antiserum. The levels of sPLA₂ measured in this modified ELISA amounted to 1.4 ± 0.25 and 32.0 ± 5.8 (mean ± standard error of mean [SEM], n = 5) ng sPLA₂/10⁷ cells, respectively, for neutrophils and eosinophils. Measurement of sPLA₂ activity as described in Materials and Methods showed an activity of 14.6 ± 2.9 (mean ± SEM, n = 3) nmol/min/10⁸ cells in sonicates from human eosinophils and 1.6 ± 0.4 (mean ± SEM, n = 3) nmol/min/10⁸ cells in sonicates from neutrophils, again indicating that eosinophils contain high amounts of sPLA₂ relative to neutrophils. It should be noted that for the sPLA₂ activity measurement, we have used crude whole cell sonicates. These sonicates might have contained endogenous inhibitors of the sPLA₂ enzyme.29 

Investigation of the intracellular localization of sPLA₂ in unstimulated eosinophils was undertaken by immunoelectron microscopy (immuno-EM). Immunogold-labeled MoAb against sPLA₂ was detected in granules that contained the typical eosinophil crystals (Fig 2). It is difficult to ascertain whether the heterogeneity in the labelling in the granules is due to heterogeneity in the localization of sPLA₂ or to partial destruction of the epitopes for the antibody due to the fixation procedure used.

Immunolocalization of sPLA₂, after addition of STZ to IL-5–primed eosinophils, was also studied with immuno-EM. After binding, STZ particles were engulfed by the eosinophils with a lag time of about 2 minutes (unpublished data). The cells were fixed 10 minutes after the addition of STZ for the preparation of sections suitable for immuno-EM. Clearly, phagosomes had been formed and immunogold particles depicting sPLA₂ were present in these compartments (Fig 3A). Some background staining was noted under these conditions (Fig 3B), but this staining was considerably lower than the staining obtained with the anti-sPLA₂ antibody.

Using the sPLA₂ ELISA, we measured the sPLA₂content in supernatants collected from unstimulated eosinophils and from eosinophils treated with PAF, fMLP, PAF-fMLP, cyto B, or cyto B-fMLP. None of these treatments induced the secretion of sPLA₂ into the supernatant. Other stimuli, such as Ca²⁺-ionophore A23187, PMA, or combinations of these stimuli, also did not result in the release of sPLA₂ in the supernatant (data not shown).

In another set of experiments, eosinophils were stimulated with STZ particles or serum-opsonized Sephadex (SOS). As shown in Table 2, ECP was released, especially after stimulation with the much larger SOS particles leading to frustrated phagocytosis (Egesten et al, manuscript in preparation). In agreement with the enhanced percentage of eosinophils able to bind opsonized particles after IL-5 pretreatment,9 the release of ECP was enhanced for both the STZ- and SOS-induced ECP release after IL-5 pretreatment. However, the same supernatants did not contain sPLA₂ as determined by the sPLA₂ ELISA (Table2).

This study has shown that human eosinophils express high amounts of the secretory 14-kD sPLA₂, whereas neutrophils and monocytes express only a small amount. Our data indicate that in future studies on the role of secretory 14-kD sPLA₂ in monocytes or neutrophils, contamination by eosinophils has to be excluded because of the very high amounts of this enzyme in eosinophils. In lymphocytes, sPLA₂ was not detectable by ELISA or by immunocytochemical staining of cytospin preparations. In basophils, a small amount could be present, the ELISA signal being on the borderline of our assay. A preliminary report has appeared describing a possible role of secretory PLA₂ in LTC₄ synthesis by human basophils.30 

Because we used for this study antibodies raised against the synovial type of sPLA₂ (belonging to group II of PLA₂s), it is most likely that the enzyme detected by us in human eosinophils is identical to this protein. However, it cannot be completely excluded that another member of the sPLA₂ gene family is highly expressed in eosinophils, eg, the recently cloned group V enzyme. Interestingly, this latter protein has recently been shown in murine macrophages19 and mast cells20 to play an important role in arachidonic acid release, a role previously assigned to the group II PLA₂ enzyme.

Although eosinophil degranulation and synthesis of PAF and LTC₄ are dependent on PLA₂ activity, the role of the (secretory) sPLA₂ in various eosinophil responses can only be speculated on. The fact that mepacrine has been suggested to specifically inhibit sPLA₂³¹ and is able to completely block STZ-induced PAF synthesis in human eosinophils9 might indicate an important role for sPLA₂ activity in eosinophil activation. In addition, the presence of sPLA₂ in the eosinophil phagosomes may indicate a function for this PLA₂ in the killing of microorganisms. Mammalian sPLA₂ is able to directly degrade phospholipids of Gram-negative bacteria.32 

Despite significant degranulation (as indicated by the release of ECP), sPLA₂ did not appear to be released from eosinophils on activation in vitro. sPLA₂ is clearly lipolytic and catalyzes a reaction at the water/lipid interface.11 Hence, it is not unlikely that directly after secretion and exposure to mmol/L Ca²⁺ concentrations, the enzyme rebinds to the membrane. However, immunoelectron microscopy did not yield an indication for the presence of sPLA₂ on the outer membrane after eosinophil activation. Likewise, we did not detect sPLA₂ on the plasma membrane by indirect immunofluorescence with the MoAbs 4A1 or 10B2 (data not shown). Possibly, insertion of the enzyme into the lipid bilayer prevents binding of the antibodies. Cell fractionation followed by detergent lysis of the various fractions may be helpful to resolve this issue. Bach et al33 have shown that quantification of release of eosinophil granular proteins is hampered by destruction on eosinophil activation. Measurement of remaining eosinophil-derived neurotoxin (EDN) and EPO in activated eosinophils showed a recovery of only 40% in the cell pellet of activated eosinophils, without apparent release of these proteins (<2%). These investigators could not explain this low recovery. Also, we have found a low recovery of sPLA₂ protein (as determined by ELISA measurements of sPLA₂ in the supernatant and the remaining amount of sPLA₂ in the cell pellet of activated eosinophils) of 51% ± 8% and 66% ± 6% (mean ± SEM, n = 3) in PAF/STZ- and PMA-stimulated human eosinophils respectively, compared with a recovery of 95% ± 5% in unstimulated eosinophils. The cause of this poor recovery is currently under investigation.

Previously, Rosenthal et al34 have shown the presence of a 14-kD PLA₂ in phagosomes of neutrophils on activation with serum opsonized S. aureus. Under certain conditions, this enzyme can be found in the supernatant of activated neutrophils.34,35 Our study shows that the level of sPLA₂ is about 20-fold higher in eosinophils than in neutrophils, but also that under in vitro conditions the enzyme is not readily released from the eosinophils. Although we were not able to show release of sPLA₂ protein from eosinophils in vitro, sPLA₂ might be released from eosinophils in vivo. Bowton et al36 have recently described enhanced sPLA₂ activity in bronchoalveolar lavage (BAL) fluid after antigen challenge in asthmatic patients. These investigators did not examine the cellular source of the sPLA₂, although they suggested the mast cell as a possible candidate. However, our results, while not refuting mast cells as an important source of sPLA₂, might suggest eosinophils as another cellular source of sPLA₂ in BAL fluids. Released sPLA₂, eg, in the lungs, may act on other cell types present in the lung: addition of sPLA₂ to mast cells results in the release of histamine, one of the agents responsible for the symptoms during an asthmatic attack.37 

The authors thank Prof. H. van den Bosch (University of Utrecht, The Netherlands) for advice on sPLA₂ activity measurements.