Essential Requirement of Cytosolic Phospholipase A2for Activation of the Phagocyte NADPH Oxidase*

Arachidonic acid (AA) can trigger activation of the phagocyte NADPH oxidase in a cell-free assay. However, a role for AA in activation of the oxidase in intact cells has not been established, nor has the AA generating enzyme critical to this process been identified. The human myeloid cell line PLB-985 was transfected to express p85 cytosolic phospholipase A2(cPLA2) antisense mRNA and stable clones were selected that lack detectable cPLA2. cPLA2-deficient PLB-985 cells differentiate similarly to control PLB-985 cells in response to retinoic acid or 1,25-dihydroxyvitamin D3, indicating that cPLA2 is not involved in the differentiation process. Neither cPLA2 nor stimulated [3H]AA release were detectable in differentiated cPLA2-deficient PLB-985 cells, demonstrating that cPLA2 is the major type of PLA2 activated in phagocytic-like cells. Despite the normal synthesis of NADPH oxidase subunits during differentiation of cPLA2-deficient PLB-985 cells, these cells fail to activate NADPH oxidase in response to a variety of soluble and particulate stimuli, but the addition of exogenous AA fully restores oxidase activity. This establishes an essential requirement of cPLA2-generated AA for activation of phagocyte NADPH oxidase.

Arachidonic acid (AA) can trigger activation of the phagocyte NADPH oxidase in a cell-free assay. However, a role for AA in activation of the oxidase in intact cells has not been established, nor has the AA generating enzyme critical to this process been identified. The human myeloid cell line PLB-985 was transfected to express p85 cytosolic phospholipase A 2 (cPLA 2 ) antisense mRNA and stable clones were selected that lack detectable cPLA 2 . cPLA 2 -deficient PLB-985 cells differentiate similarly to control PLB-985 cells in response to retinoic acid or 1,25-dihydroxyvitamin D 3 , indicating that cPLA 2 is not involved in the differentiation process. Neither cPLA 2 nor stimulated [ 3 H]AA release were detectable in differentiated cPLA 2 -deficient PLB-985 cells, demonstrating that cPLA 2 is the major type of PLA 2 activated in phagocytic-like cells. Despite the normal synthesis of NADPH oxidase subunits during differentiation of cPLA 2 -deficient PLB-985 cells, these cells fail to activate NADPH oxidase in response to a variety of soluble and particulate stimuli, but the addition of exogenous AA fully restores oxidase activity. This establishes an essential requirement of cPLA 2 -generated AA for activation of phagocyte NADPH oxidase.
The phagocyte NADPH oxidase is a multicomponent transport chain that transfers electrons from NADPH to molecular oxygen and generates superoxide, a precursor of microbicidal oxidants important to host defense. NADPH oxidase subunits include three cytoplasmic components, p47 phox , p67 phox , and Rac-2, and a membrane flavocytochrome b 558 composed of gp91 phox and p22 phox (1)(2)(3)(4)(5)(6)(7)(8)(9). In differentiated phagocytic cells stimulation results in translocation of the cytosolic NADPH oxidase components to the membrane where they interact with the flavocytochrome to form the activated oxidase leading to superoxide generation. The signals responsible for assembly and activation of the oxidase are not clearly defined. PLA 2 1 activity has been implicated in a variety of responses by stimulated phagocytes, including degranulation, phagocytosis, adhesion, cell spreading, and activation of NADPH oxidase (10 -16). Until recently, generation of AA had been viewed as a modulating event leading to oxidase activation but not as crit-ical to the process. Recently we and others have used PLA 2 inhibitors to implicate generation of AA as important for activation of the NADPH oxidase activity in human neutrophils (17,18). Moreover, we have shown that AA increases the affinity of the assembled oxidase for NADPH (19). However, studies using inhibitors may not delineate the specific enzyme of a related group whose inhibition is responsible for the observed effect, and the inhibition seen may result from action on more than one enzyme.
In the last decade, several secreted and cytosolic mammalian PLA 2 isozymes have been described (20 -24). The existence of several types of PLA 2 in phagocytic cells (25)(26)(27)(28)(29)(30) complicates delineation of the PLA 2 responsible for release of the AA, which impacts on NADPH oxidase activation following phagocyte stimulation. In the present study, we used the RNA antisense technique to create in the human phagocyte myeloid cell line, PLB-985, a p85 cPLA 2 -deficient model cell line to demonstrate the role of this enzyme in NADPH oxidase activation.
Construction of Expression Vector-Human cPLA 2 cDNA (in a PMT2 vector) was generously provided by Dr. J. L. Knopf (Genetic Institute, Cambridge, MA). cPLA 2 cDNA (1-530) was excised from its PMT2 vector with the restriction enzymes SalI and XbaI. The ends were filled in with Klenow fragment and ligated in the antisense direction into the XhoI site of pcDNA3 (Invitrogen, San Diego, CA), which was also filled in with Klenow to form the plasmid cPLA 2 (1-530)-pcDNA3. The antisense orientation of this insert was confirmed by DNA sequence analysis.
Transfection and Selection of PLB-985 Clones-PLB-985 cells (1 ϫ 10 7 ) in logarithmic growth were transfected in 0.3 ml of culture medium with 20 g of plasmid DNA (antisense or vector alone) by electroporation at 250 V and 960 microfarad in a Gene pulser Unit (Bio-Rad, Melville, NY) and selected in the presence of 0.8 mg/ml G418 (Life Technologies, Inc.) (33). At 48 h. posttransfection, the living cells were separated on a Ficoll gradient (typically 30 -50% of starting number), diluted to 5 ϫ 10 4 cells/ml, and plated in 96-well plates in the presence of the appropriate antibiotic for selection of clones (G418 (Life Technologies, Inc.) 0.8 mg/ml). Clones resistant to the neomycin analogue, G418, were screened by Western blot to select those that were cPLA 2 protein-deficient.
Immunoblot Analysis-For immunoblot detection of cPLA 2 , cells were centrifuged and sonicated in the presence of 5 mM EGTA, 100 mM KCl, 3 mM NaCl, 3.5 mM MgCl 2 , 10 mM PIPES (pH 7.4), 1 mM PMSF, and 100 M leupeptin, and membrane and cytosol fractions were separated as described (34). Under these conditions whole cell cPLA 2 protein and activity were found entirely in the cytosol fraction as shown earlier (35). Alternatively, cell lysates were prepared using 1% Triton X-100, 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 10% * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
DTT-resistant Phospholipase A 2 Activity-PLB-985 cells (5 ϫ 10 6 cells/ml) in HBSS buffer were stimulated for 2 min at 37°C by 1 mg/ml OZ or 50 ng/ml PMA. The reaction was stopped by 10-fold dilution with cold HBSS and immediate centrifugation at 4°C. Membrane and cytosol fractions were separated as described earlier (35). PLA 2 activity was determined in the cytosols using sonicated dispersions of 1-stearoyl-2-[ 14 C]arachidonyl phosphatidyl choline (30 M, 50,000 dpm/assay) and sn-1,2-dioleoylglycerol (molar ratio, 2:1) in an assay mixture containing 5 mM DTT (36, 37) with some modifications as described earlier (35). Briefly, the assay mixture contained the phospholipid substrate in 80 mM KCl, 5 mM CaCl 2 , 5 mM DTT, 1 mg/ml bovine serum albumin, 1 mM EDTA, and 10 mM HEPES (pH 7.4). The reaction was started by the addition of 50 g of cytosolic protein (within the linear protein range of the assay) and incubated at 37°C in a shaking water bath for 10 min. Mac-1 antigen determination was detected by indirect immunofluorescence as described earlier (38).
Superoxide Anion Measurements-The production of superoxide anion (O 2 . ) by intact cells was measured as the superoxide dismutase inhibitable reduction of ferricytochrome c (34). Cells were suspended (5 ϫ 10 5 cells/well) in 100 l of HBSS containing 150 M ferricytochrome c and activated by the addition of the appropriate stimulus. The reduction of ferricytochrome c was followed by a change of absorbance at 550 nm every 2 min over a 20-min time course using a Thermomax microplate reader (Molecular Devices, Menlo Park, CA). The maximal rates of superoxide generation were determined using a extinction coefficient E 550 ϭ 21 mM Ϫ1 cm Ϫ1 . OZ was prepared as follows: 20 mg zymosan was incubated with 1 ml of pooled human serum (lipopolysaccharide-free) for 1 h at 37°C and washed three times with HBSS.

RESULTS AND DISCUSSION
PLB-985 cells were transfected with human cPLA 2 cDNA engineered in the antisense orientation in the pcDNA3 expression vector, which also contains a neomycin resistance element. Cell cytosolic fractions from clones resistant to the neomycin analogue, G418, were screened by Western blot to select those that were cPLA 2 protein-deficient. Three pcDNA3-cPLA 2 antisense transfected clones were completely deficient of cPLA 2 protein (PLB-D) (Fig. 1A). Several other clones were shown to be partially deficient, containing residual amounts of cPLA 2 protein (PLB-R). Both the parent PLB-985 line (PLB) and a G418-resistant clone transfected with the empty pcDNA3 vector (PLB-V) were used as controls. As shown in Fig. 1A, PLB-V produced high levels of cPLA 2 protein, and this amount was indistinguishable from that present in the parent PLB cells.
Cytosol fractions of the various PLB-derived cell lines were examined for the presence of DTT-resistant enzymatic activity characteristic of p85 cPLA 2 using 1-stearoyl-2-[1-14 C]arachidonyl phosphatidylcholine as a substrate. As shown in Fig. 1B, basal cPLA 2 activity resistant to inhibition by 5 mM DTT detected in unactivated PLB-V cytosol was similar to that seen in the parent PLB cells (1.25 Ϯ 0.2 and 1.46 Ϯ 0.3 pmol/mg protein/min, respectively). By contrast PLB-D cells that lack cPLA 2 protein had no detectable DTT-resistant cPLA 2 activity, and PLB-R that express reduced levels of cPLA 2 protein demonstrated reduced levels of cPLA 2 activity in cytosol (0.71 Ϯ 0.4 pmol/mg protein/min). In other experiments (not shown), the cPLA 2 activity and protein detected in Triton-extracted cell lysates were identical to that observed in the cytosol fractions shown in Fig. 1.
An important characteristic of the PLB cell line is that it can be induced to differentiate toward a mature phagocyte phenotype in response to a variety of differentiating agents (31). In the present study, 10 Ϫ6 M RA was used to induce maturation to a granulocyte-like phenotype, whereas 5 ϫ 10 Ϫ8 M 1,25-dihydroxyvitamin D 3 (D 3 ) was used to induce maturation to a monocyte-like phenotype (34). Of importance for the current study is that common features of the differentiated phenotype acquired during differentiation with both induction regimens included expression of CD11b (Mac-1 antigen) at the cell surface and expression of subunits required for assembly of active NADPH oxidase. As shown in Fig. 2A, the percentage of parent PLB cells expressing Mac-1 antigen at the cell surface measured by indirect immunofluorescence was less than 8% in undifferentiated cells but increased to over 70% by day 5 with either RA or D 3 . Of note is that the pattern of Mac differentiation was indistinguishable from that seen with PLB cells. Similarly, the patterns of expression of the subunit components of NADPH oxidase during differentiation of PLB-V, PLB-D, or PLB cells in response to RA or D 3 were indistinguishable, as shown by the Western blots in Fig. 2 (B and C). Rac-2 was detected in all the PLB-derived cell lines and did not change during differentiation (not shown). Because cPLA 2 could be important for cell growth, we examined DNA synthesis and proliferation using [ 3 H]thymidine incorporation and serial cell counts, respectively. We found that parent PLB, PLB-V, and PLB-D had similar [ 3 H]thymidine incorporation and cell doubling times in the uninduced state and following differentiation showed a similar decline in these measures as maturation occurred (not shown). These analyses of phenotype changes during differentiation indicate that neither the presence of plasmid sequence (PLB-V) nor loss of cPLA 2 (PLB-D) had any effect on proliferation or differentiation-associated expression of Mac-1 antigen or NADPH oxidase components.
Differentiation was not associated with any change in the constitutive level of cPLA 2 protein (Fig. 3A) or enzymatic activity in cytosol fractions (Fig. 3B) during maturation of parent PLB or PLB-V cells. These results are expected because a recent report (39) indicates that constitutive expression of cPLA 2 does not change during differentiation of HL60 cells, a human promyelocytic cell line. PLB-D remained devoid of measurable cPLA 2 protein or activity following differentiation (Fig. 3).
As shown in Table I, we compared activation of the NADPH oxidase in differentiated PLB, PLB-V, PLB-R, and PLB-D in response to both soluble and particulate stimuli. PLB and PLB-V show similar levels of superoxide generation when similar differentiation conditions and stimulants are compared. The differentiated PLB-R, which expresses some cPLA 2 but less than that seen in differentiated PLB or PLB-V, generated superoxide at levels 50 -60% of that seen for differentiated PLB or PLB-V for each stimulus. Differentiated PLB-D did not generate superoxide in response to any of the stimuli tested despite the fact that there were normal levels of expression of the NADPH oxidase components. Thus, the level of cPLA 2 expression appeared to have a profound effect on NADPH oxidase generation of superoxide.
Because cPLA 2 would most likely mediate its action on superoxide production by the NADPH oxidase in stimulated cells by release of AA, its role in the liberation of AA was studied. We prelabeled undifferentiated and 5 day D 3 -differentiated PLB with [ 3 H]AA and then measured release of [ 3 H]AA following stimulation with PMA or OZ (Fig. 4A). response to PMA or OZ relative to that seen with the similarly stimulated undifferentiated cells. The differentiated PLB-D showed a striking difference in that the level of stimulated [ 3 H]AA release remained at the same low level as that seen in undifferentiated cells. This suggests that in differentiated phagocytes the major portion of the pulse of AA release associated with PMA or OZ stimulation is mediated by cPLA 2 . It also suggests that the failure of differentiated PLB-D to produce superoxide is related to a failure to produce this large pulse of AA release following stimulation.
We tested this latter hypothesis by providing exogenous free AA to D 3 -differentiated PLB-D at the time of stimulation of the oxidase by PMA as analyzed in the kinetic assay of superoxide generation shown in Fig. 4B. Addition of 25 MAA to the differentiated PLB-D fully restored the superoxide generating capacity to these cells relative to rates of superoxide generation by parent PLB cells stimulated with PMA. Not shown is that addition of free AA alone in the range of 10 -25 M without PMA did not induce any superoxide production. Furthermore, the effect of AA was shown to be specific, because the addition of linoleic or oleic acids (up to 100 M) to PMA-stimulated PLB-D cells did not restore oxidase activity. This demonstrates that it is the AA generated by cPLA 2 that is essential to superoxide generation by the activated NADPH oxidase.
NADPH oxidase activation requires translocation of the cytosolic subunits to the cell membrane following stimulation (1)(2)(3)(4)(5)(6)(7)(8)(9). As shown in the Western blot in Fig. 4C, stimulation of differentiated PLB-D resulted in plasma membrane association by the three cytoplasmic components of the NADPH oxidase that was indistinguishable from that seen in differentiated PLB-V cells. It is of note that addition of AA to differentiated PLB-D at the time of PMA stimulation did not augment the normal levels of cytosolic NADPH oxidase subunits associated with the plasma membrane (Fig. 4C, last two lanes). This suggests that AA is acting in some final step on the assembled oxidase to promote or maintain the enzyme in an active state rather than acting to induce assembly of the oxidase. Such a model is consistent with previous observations showing that PLA 2 inhibitors do not affect p47 phox membrane translocation or phosphorylation (17) and that PLA 2 inhibitors work even when added after oxidase activation (18).
Our studies are the first to delineate the effects of complete and specific inhibition of p85 cPLA 2 protein expression and strongly implicates the p85 cPLA 2 as the PLA 2 isozyme responsible for the majority of the pulsed release of AA following stimulation of phagocytes. It also provides the first unequivocal  demonstration that AA release is an absolute requirement for superoxide generation by the NADPH oxidase regardless of the stimulus. A number of previous studies have suggested that AA release in stimulated neutrophils or monocytes is mediated by p85 cPLA 2 and that this AA release is important for a variety of functional responses. Specifically, it has been shown that the decreased expression of cPLA 2 in monocytes treated with antisense cPLA 2 oligonucleotide resulted in significant inhibition of [ 3 H]AA release induced by monocyte chemotactic protein-1 (40). Furthermore, treatment of human monocytes with bacterial lipopolysaccharide caused an increase in the 85-kDa PLA 2 protein and activity that coincided with increased release of PGE 2 (41). This increase in PGE 2 formation could be prevented by treatment of the monocytes with an antisense oligonucleotide directed at the 85-kDa PLA 2 mRNA, which decreased cPLA 2 levels. Bacterial lipopolysaccharide primes human neutrophils for enhanced release of AA and causes phosphorylation of cPLA 2 (42). In monocytes a correlation has been demonstrated between monocyte colony-stimulating factor-induced phosphorylation events, increased expression of 85-kDa PLA 2 , and increased PGE 2 release (43). It was recently suggested (44) that cPLA 2 is the enzyme that induces the release of AA in a permeabilized neutrophil model.
In conclusion, the development of differentiated PLB-D cell lines that lack any cPLA 2 expression demonstrates that this protein is not required for proliferation or differentiation processes but that its enzymatic production of AA is an essential requirement for activation of the phagocyte NADPH oxidase. Although AA has been used extensively to activate the NADPH oxidase in a cell-free system (34), it was not known whether its effect in vitro had a correlate in the intact cell. Because our previous studies suggest that AA markedly enhances the affinity of the NADPH oxidase for the NADPH substrate (19), our current observations are most consistent with a model in which cPLA 2 -generated AA might be a co-factor acting in the intact phagocytic cell to enhance the affinity of the assembled NADPH oxidase for NADPH.