Inhibition of the Oxidative Burst in Human Neutrophils by Sphingoid Long-chain Bases ROLE OF PROTEIN KINASE C IN ACT~ATION OF THE BURSF

The neutrophil oxidative burst is characterized by increased ceIluIar 0 2 consumption due to the activation of a membrane-associated superoxide-generating NADPH-oxidase. The response is triggered by a variety of stimuli, including opsonized zymosan, formyl- methiony~leucinephenyl~anine (FMLP), arachidonate, short-chain diacylglycerols, and phorbol myris- tate acetate (PMA). We herein demonstrate that incubation of cells with sphinganine or sphingosine blocks or reverses activation by these agonists. The inhibition is reversible, does not affect cell viability, and does not affect another complex cell function, phagocytosis. Inhibitory c o n ~ n t r a t i o ~ of sphznganine did not signif- icantly affect cytoplasmic calcium levels or FMLP-generated calcium transients. Structural requirements for inhibition of the oxidative burst include a long aliphatic chain and an ~ino-containing head-group, and there is modest specificity for the native (eryttrro) isomer of sphinganine. Inhibition involves stimulus- induced activation mechanisms rather than a direct effect on the NADPH oxidase, since sphinganine did not inhibit NADPH-dependent superoxide generation in isolated membranes containing the active enzyme. Activation by FMLP, diacylglycerol,

membrane to form an enclosed vacuole, the phagosome. Various bactericidal mechanisms are then triggered including the "oxidative burst" (1, 2) which involves a marked increase in cyanide-insensitive 0, consumption, with generation of superoxide, hydrogen peroxide, and hydroxyl radical (see Ref. 3 for review). The first step in production of these species involves the NADPH-dependent reduction of oxygen to generate superoxide (0;) (3,4), catalyzed by the NADPH oxidase (5,6) which appears to be in both the plasma membrane and the specific granules (7,8).
The NADPH oxidase is dormant in the resting neutrophil. When exposed to a variety of either particulate or soluble stimuli the oxidative burst is triggered apparently by activation of the oxidase (3,4). Phagocytosis of bacteria, opsonized latex, or opsonized zymosan (immunoglobulin-treated yeast cell walls) activates the burst (9). Likewise, chemotactic agents such as formylmethionyl~eucinepheny~aianine (FMLP') and C5a activate via cell surface receptor mechanisms (10,ll). The tumor promotor phorbol myristate acetate (PMA) and the lipid ~~t a n o l y g l y c e~l , both known activators of protein kinase C also stimulate (11,12), as do arachidonate and other unsaturated fatty acids (13). Calcium has been implicated, but its role remains controversial; two calcium ionophores (A23187 and ionomycin) provide conflicting results (14). Although there has been a great deal of recent interest in the mode of activation by various stimuli, the intracellular messengers and mechanisms by which these agonists produce a cellular response remains elusive. Apparently, all activators ultimately stimulate the same NADPH oxidase (15). However, several mediation mechanisms have been proposed, and evidence suggests that not all stimuli share a common pathway. Nevertheless, it is possible that several or all initial pathways converge in a final sequence of regulatory events. Proposed messengers and components of signal mediation pathways include arachidonate (13,16,17), calcium and calmodulin (18,19), protein kinase C (11,201, and cAMP/cAMP-dependent kinases (21,229 (see McPhail and Snyderman (23) for a recent review).
The lipids sphinganine and sphingosine are shown in the accompanying paper (24) to be inhibitors of protein kinase C in a mixed mice~e-r~onstituted system. Because protein kinase C has been implicated in at least some neutrophil oxidative burst activation mechanisms, we undertook the present studies, first to test the regulatory effects of these lipids in this cellular system, and second to use the lipids as probes of intracellular regulatory mechanisms. We find that The abbreviations used a r e : FMLP, formylmethionylleucinephenylalanine; PMA, phorbol myristate acetate, BSA, bovine serum albumin; ACTH, adrenocorticotropic hormone.

12616
This is an Open Access article under the CC BY license.
long-chain bases are potent inhibitors of the neutrophil oxidative burst and that a variety of agonists appear to share at least one common step in their activation pathways. We also provide evidence that this step is likely to involve the action of protein kinase C.

RESULTS
Inhibition by Sphinganine of 0, Consumption a d Superoxide-mediated Reduction of Cytochrome c-Sphinganine was tested for its ability to inhibit the oxidative burst of normal human neutrophils. This lipid rather than sphingosine was used initially as the prototypical long-chain base, since it is commercially available in synthetic chemically well characterized form. When neutrophils were preincubated with sphinganine, activation of O2 consumption by PMA was abolished (see Fig. 1, Panel A). The increased oxygen consumption is thought to reflect reduction of O2 to superoxide anion, which can be detected by monitoring superoxide dismutaseinhibitable cytochrome c reduction. Fig. 1, Panel B, shows that preincubation with sphinganine also blocks PMA activation of superoxide generation. Fig. 1 (Panel A, Tracing 3) also illustrates that the addition of sphinganine several minutes after initiation of the oxidative burst causes a gradual return to the basal rate of oxygen consumption. (The apparent loss of activity at later times in the control tracing (PMA stimulated, no sphinganine addition) was due to oxygen depletion, in contrast to the true inactivation seen following sphinganine addition.) The decay to basal rates followed first order kinetics as determined by plotting the natural log of the rate of oxygen consumption after addition of sphinganine as a function of time (not shown). The half-time for inactivation varied from 30 s to 6 min with the individual donors examined. Since sphinganine added 2-3 min prior to activation with PMA completely abolished the oxidative burst for all donors, the inactivation rate does not appear to reflect the rate at which sphinganine gains access to the cell. Rather, we suggest that the variable rate of this decay represents individual variation in a normal inactivation process which is revealed when activation mechanisms are blocked.

E~idence That S p h~~a n~n e Acts ~e v e r s i~~ and w i t~u t Cytotoxicity on the NADPH Oxidase Activation Mechanisms-
The inhibitory effects of sphinganine were not due to cytotoxic effects of long-chain bases (38). When neutrophils (1.2 X lo7 cellsfml) were incubated with 50 HM sphinganine, a completely inhibitory concentration, there was no effect on cell viability. Trypan blue exclusion showed greater than 95% viability, and lactate dehy~ogenase release showed greater than 96% of cellular lactate dehydrogenase was retained in both the presence and absence of sphinganine.
To determine whether the sphinganine inhibition was reversible, neutrophils (1.2 X lo7 cells/ml) were first incubated  phate-buffered saline/glucose containing 50 gM fatty acidfree BSA. After such procedures, restoration of PMA-induced oxygen consumption comparable to that in untreated cells was seen (3.7 and 3.8 nmol/min/106 cells, respectively), indicating that inhibition of the oxidative burst by s p h i n g~i n e is reversible.
To determine whether the observed effects of sphinganine were on the activation process or on the NADPH-dependent oxidase enzyme itself, membranes were prepared from PMApretreated cells, and examined for sphinganine effects on superoxide generation. Membranes isolated from PMA-pretreated cells have previously been shown to exhibit an approximately 10-fold increase in NADPH-dependent 0; generation. No inhibition was observed when these membranes were incubated with concentrations of sphinganine which inhibited the cellular response (see Table I). In fact, a slight stimulation was seen, as is also seen with some detergents (39). Thus, sphinganine does not inhibit the oxidase enzyme directly, but must act at some site(s) involved in activation. Burst Induced by a Variety of Activators-The inhibitory effect of sphinganine was tested using a variety of activators. Table I1 shows that with all activators tested there was complete inhibition of the oxidative burst by sphinganine. To test whether all stimulants utilized a common, sphinganine-inhibitable step, the concentration dependence for inhibition was determined for a variety of activators. In initial studies with dioctanoylglycerol and PMA, oxygen consumption was monitored (Fig. 2, Panel A).

Lack of Effects of Sphinganine on
Using this assay, half-maximal inhibition occurred at 7.4,6.7, and 5.6 y~ using dioctanoylglycerol, PMA, and opsonized zymosan, respectively. These values are identical within experimental error and individual variation among donors. Thus, as has been suggested by a variety of earlier studies (e.g. see Ref. 20), phorbol esters and diacylglycerols appear to activate by a common mechanism. In a d~t i o n , activation by opsonized zymosan appears to share this sphinganine-inhibitable mechanism. Data are expressed plus or minus the standard deviation of the mean.

(5 pelml)
TO quantitate sphinganine inhibition using a series of other activators, the superoxide assay was used (Fig. 2, Panel B). This assay is considerably more sensitive and less time consuming than the oxygen consumption assay, therefore requiring fewer cells and facilitating data acquisition. Activation by both FMLP and arachidonate was 50% inhibited at essentially the same concentration as that which affected PMA activation, again implicating a common site of inhibition. (Particulate activators could not be assessed using this assay due to turbidity and base line noise.) There is an apparent but readily explainable discrepancy in sphinganine concentration required for 50% inhibition using the two assays (compare, for example PMA activation in Panel A versus Panel B of Fig. 2). This is due to the difference in cell number used in the two assays. The oxygen consumption assay requires large numbers of cells (approximately 7.5 x lo6 cells/ml) and shows a concentration dependence for 50% inhibition of 6-8 p M sphinganine. In separate experiments (not shown) we found that inhibition of oxygen consumption required higher sphinganine concentration when the cell number was increased and lower when the cell number was decreased. The superoxide production measurements were obtained using about 10-fold less cells (approximately 5 X lo6 cells/ml), and required about 1 PM sphinganine for 50% inhibition for all activators. Thus the difference in required concentration can be readily explained by a cell number effect. The concentration requirement appears to be approximately directly proportional to cell number as is also seen for many lipophilic compounds (e.g. diacylglycerols and arachidonate) whose activity requires partitioning into cellular membranes. This is in agreement with sphinganine inhibition being subject to surface dilution effects in an in vitro mixed micelle system and in platelets (24).
To determine how much inhibitor was actually becoming associated with the cells, radiolabeled sphinganine (3, 10, 30 p~) plus albumin carrier was incubated with neutrophils (2.5 X lo6 cells) for 5 min (i.e. a time longer than that needed to achieve inhibition). After washing by centrifugation, the cellassociated radioactivity was measured. Approximately 20% of the added sphinganine became cell-associated. Therefore, because of partitioning between the albumin and the cells, the cellular concentration of sphinganine was considerably less than the total. Thus, if generated intracellularly in the appropriate compartment, long-chain bases are expected to be more effective than the concentration dependence for externally added inhibitor indicates. Effect of Structural Analogs of Sphinganine on the Oxidative Burst-Using PMA activation, analogs with structural features similar to those of sphinganine were tested for their effects on the oxidative burst (see Fig. 3 and Table 111). Compounds which also showed inhibitory effects include stearylamine, sphingosine, threo-sphinganine, and octylamine. Other compounds tested (see Table 111) did not show inhibition at 100 pM concentrations.
It appears that the structural features necessary for inhibition include a free amino group and a long alkyl chain. There is also a modest selectivity for the native stereoisomer, erythro-sphinganine, since the threo-isomer was 3-4 times less effective. The longer alkyl chain derivatives were more effective than shorter, and a hydrophobic benzene ring could not substitute for an alkyl chain. Thus, sphinganine and  sphingosine were the most effective inhibitors tested, and inhibition appears to exhibit some degree of structural specificity.
Effects of Sphinganine on PMA-induced Phosphorylation-PMA has been previously shown to induce phosphorylation of up to 6 proteins in rabbit neutrophils (40). Table IV shows the effects of sphinganine on trichloroacetic acid-precipitable counts of 3zP in PMA-treated cells. The sphinganine/PMAtreated cells show essentially the same counts as both the control and sphinganine-treated cells. The PMA-treated cells showed approximately double the number of counts. Fig. 4 shows that PMA also stimulates phosphorylation of a large number of proteins in human neutrophils (compare lanes A and C). Although it is difficult to resolve all individual proteins in one-dimensional gels, there is a definite PMA-induced increase in 4-5 bands of molecular weights similar to those documented previously (40), plus a general increased density of phosphorylation in the entire lane. Fig. 4 shows also that a concentration of sphinganine which inhibits the oxidative burst blocks this PMA-induced phosphorylation. (Long-chain base also appeared (see lanes C and D ) to increase the phosphorylation of 1-2 bands. Although this possibility is under continuing investigation, the increased phosphorylation in this area did not appear to be a consistent feature.) Since PMA binds to and activates protein kinase C, resulting in the observed phosphorylation, the inhibition by sphinganine also appears to be a direct effect on protein kinase C.
Sphinganine Inhibition of Radiolabeled Phorbol Dibutyrate Binding-Displacement of bound, radiolabeled phorbol dibutyrate by diacylglycerols has been used previously (41) to provide evidence for a common cellular site of action (protein kinase C) for these two activators. Herein, we have used the same technique to evaluate whether, as indicated by phosphorylation studies, sphinganine is acting by binding to protein

Protein Kinase C
Control of the Neutrophil Oxidative Burst kinase C. Fig. 5 shows that sphinganine displaces phorbol dibutyrate from its cellular binding site. Control studies were also carried out using 50 p~ ceramide and palmitic acid, two analogs which did not show significant inhibition (see Table  111). With these analogs there was 18% and 3% displacement, respectively, compared with 60% displacement with the same concentration of sphinganine. Residual apparent binding is nonspecific, since an excess of unlabeled PMA produced the same degree of displacement. These studies show that the decrease in binding of the phorbol ester does not appear to be due to detergent or other effects. Rather, there appears to be a specific inhibition of the phorbol ester binding by sphinganine. Such behavior was also seen in the micelle-reconstituted system and in human platelets (24).
The half-maximal inhibition of the phorbol dibutyrate (50 nM)-activated oxidative burst (superoxide generation) occurred at approximately 0.5-1 p~ (data not shown), compared with 20 p~ in the binding displacement experiment. However, when the binding experiment was carried out with the same time course (i.e. 5 min total incubation time rather than 25 min), the half-maximal concentration for displacement was 0.7 pM, in excellent agreement with the activity effects. We speculate that during the longer time course of the binding experiment shown in Fig. 5, there has been either significant metabolism of the added long-chain base, or partitioning into additional cellular compartments. Either, or both of these processes operating together, would be expected to shift the displacement curve towards a higher concentration range.
Lack of Effect of Sphinganine on Cytoplasmic Calcium and on Calcium Transients-One additional possible mechanism by which long-chain bases might act is on calcium levels and/ or on agonist-induced changes in cytoplasmic calcium concentration. FMLP is known (41) to produce a rapid (seconds) rise in neutrophil cytosolic calcium levels, followed by a gradual (about 5-10 min) return towards resting levels. Using the Quin2 assay (see "Experimental Procedures"), fluorescence was monitored continuously before and after addition of FMLP, and was used to calculate (37) intracellular calcium concentrations. Using this system PMA had no effect on   0.20 -t 0.02 (3) 0.66 & 0.14 (3) "Intracellular calcium was measured using Quin2, as described under "Experimental Procedures." Each measurement utilized 7.5 X 10' cells in 2 ml of buffer. The "FMLP peak" value refers to the maximum calcium concentration achieved following 1 p~ FMLP addition, but prior to the gradual return of calcium towards resting levels. Numbers in parenthesis refer to the number of experiments averaged to obtain the reported values. Fluorescence was corrected in the sphinganine experiments for a small amount of fluorescence of the albumin carrier. calcium levels. Table V summarizes resting and FMLP-induced "peak" calcium levels in the absence and presence of sphinganine. Sphinganine does not appear significantly to affect either the resting calcium level or the FMLP-induced rise in calcium. In addition, the kinetics of return of calcium towards resting levels (not shown) were identical. Thus sphinganine does not appear to exert its regulatory effects by modulating changes in these levels. In addition, these experiments provide another sensitive index of cellular integrity. Since the extracellular buffer contained 1 mM calcium, any loss of membrane integrity would have been reflected as a marked increase in cellular calcium.

Long-chain Bases as Possible Intracellular Regulatory
Molecules-Lipids and lipid metabolites appear to play important roles as intracellular signal molecules. For example, bioregulatory roles for diacylglycerol, arachidonate (and arachidonate metabolites such as prostaglandins), inositol triphosphate (derived from phosphatidylinositol diphosphate), and platelet-activating factor (l-0-alkyl-2-acetyl-sn-glycerol-3-phosphatidylcholine) have been documented (42-45). Hannun et al. (24) discovered that long-chain bases inhibit protein kinase C in a micelle-reconstituted system. In the present studies, the lipids sphinganine and sphingosine are shown to be potent inhibitors of the neutrophil oxidative burst. These long-chain bases are documented in accompanying papers as regulators of cell growth and differentiation in the human promyeloid leukemia cell line, HL-60 (46), and as effectors of protein kinase C in human platelets (24).
Sphinganine, a sphingoid long-chain base, is the product of the enzymatic condensation of palmitoyl-CoA and serine by serine palmitoyltransferase. Sphinganine can be incorporated into a variety of more complex biomolecules, the sphingolipids (e.g. sphingomyelin, ceramide, and glycosphingolipids). Sphingosine is also an intermediate in sphingolipid biosynthesis, can be generated from the breakdown of sphingolipids, and is the predominant long-chain base in vivo. The present studies provide support for the proposal that one or more of the long-chain bases is an intracellelar regulatory molecule. Although it remains to be shown whether these molecules are synthesized or mobilized in response to physiologic stimuli, conditions function as "anti-tumor promotors." Regulation of the Neutrophil NADPH Oxidase: Implications from Sphinganine Inhibition-The inhibition by long-chain bases of the oxidative burst appears to be a specific rather than a generalized metabolic effect, Incubation with sphinganine for the time required for the assays had no effect on cell viability and did not affect another integrative cell function, phagocytosis. Furthermore, sphinganine had no direct inhibitory effect on the NADPH oxidase enzyme system. We therefore conclude that the inhibition of the oxidative burst involves the activation process rather than the oxidase itself.
That the sphinganine does not have generalized effects on a variety of regulatory systems is indicated by at least three observations. First, in a system known to be CAMP-dependent, ACTH stimulation of steroidogenesis in the Y-1 tumor cell line, sphinganine did not affect the ACTH-stimulated steroidogenic rate.4 Second, sphinganine does not appear to affect Ca2+/calmodulin-dependent protein kinase, as evidenced by lack of inhibition of phosphorylation of the 20-kDa protein in platelets (24) using concentrations which completely abolished phosphorylation of the 40-kDa protein known to be phosphorylated by the protein kinase C (51). Third, long-chain bases do not significantly affect resting or FMLP-stimulated calcium levels. FMLP is thought to influence calcium levels initially via receptor-mediated activation o f a phospholipase which cleaves phosphatidy~inositol 45bisphosphate into diacylglycerol plus inositol trisphosphate (48)(49)(50). The latter promotes release of intracellular calcium stores. Thus, it appears unlikely that long-chain bases exert their effects on the phospholipase or on calcium levels.
Two lines of evidence from the present studies implicate protein kinase C as a specific site of inhibition by long-chain bases. First, we have shown that sphinganine can displace ["Hlphorbol dibutyrate from the phorbol ester binding site. Phorbol esters have been shown to bind to and activate protein kinase C (20). According to data provided by Bell and co-workers in the accompanying manuscript (24), sphinganine and phorbol esters compete for binding to protein kinase C.
Second, PMA-stimulated phosphorylation is inhibited by sphinganine. Thus, long-chain bases appear specifically to block activation of protein kinase C. It is possible that sphinganine may have additional as yet unknown modulatory effects on other enzymatic systems. Nevertheless, the combination of data presented herein and in the accompanying papers (24, 46) strongly implicates protein kinase C as the affected site.
The mechanism(s) of activation of the oxidative burst by what appears to be a diverse array of stimuli is (are) poorly understood. We have utilized sphinganine as a general probe of these mechanism(s~ in human neutrophils. More specifically, we wished to determine whether protein kinase C participates in some or all of the activation processes. Probably the best understood of the chemical activators are the phorbol esters such as PMA, which are potent activators of the enzyme (20). Also, synthetic diacylglycerols have been shown to be direct activators of protein kinase C (12,51). Both of these compounds appear to activate by binding to protein kinase C. In this study, it has been shown that activation of the oxidative burst by both compounds is inhibited by identical concentrations of sphinganine.
More complex in its mechanism, the chemotactic peptide FMLP functions by first binding to a plasma membrane surface receptor, thus promoting the hydrolysis of phosphatidylinositol 4,5-bisphosphate in a pertussis toxin-inhibitable process (52). Furthermore, stimulation with FMLP causes an ' E. Wilson, unpublished studies.
increase in the concentrations of both calcium and diacylglyc-erolf48). Thus, a direct pathway for protein kinase C invoivement in oxidative burst can be postulated for FMLP activation via the generated diacylglycerol, but an elevation in cytosolic calcium has also been proposed to mediate the effects. In the present studies, sphinganine inhibited the oxidative burst without affecting the cytoplasmic FMLPdependent rise in calcium concentrations. Thus a direct mediator role for calcium in activation of the oxidative burst seems unlikely. In measurements of Op production, we have shown that the half-maximal concentration of sphinganine required to inhibit superoxide production was virtually the same when either FMLP or PMA was used as the activator implicating a common inhibition site (;.e. protein kinase C).
Likewise, activation of the oxidative burst by opsonized particles (e.g. zymosan) is inhibited by sphinganine in the same concentration range which inhibited PMA activation. We conjecture that concurrent with phagocytosis, a protein kinase C activator (e.g. diacylglycerol or arachidonate) is generated, resulting in the observed oxidative burst. Thus, while phagocytosis itself is not inhibited by sphinganine the concurrent activation of the oxidative burst is blocked, consistent with the involvement of protein kinase C in this pathway. The sphinganine-inhibited cells may therefore provide a model for some forms of chronic ~a n u~o m a t o u s disease, a condition in which the patient's neutrophils phagocytose microorganisms normally, but fail to mount an oxidative burst, thus resulting in difficulty in fighting infection.
Finally, we have studied the effect of sphinganine on activation of superoxide generation by arachidonate. The mechanism by which arachidonate acts is controversial. It has previously been shown that exposure of neutrophils to opsonized zymosan, calcium ionophore, and chemotactic factors triggers the release of arachidonate. Badwey et ai. (53) showed that cis-unsaturated fatty acids (e.g. arachidonate and linoleate) activate the generation of superoxide, and produce changes in cell morphology. Inhibitors of cyclooxygenase (involved in the conversion of arachidonate to prostaglandins] and lipoxygenase (involved in leukotriene synthesis) did not affect the ability of added arachidonate to stimulate the oxidative burst, indicating that the activator is arachidonate itself rather than a metabolite. Two laboratories reported simultaneously that arachidonate activates the oxidative burst in cell-free systems (54,551. Arachidonate acts without the characteristic lag of other agonsits including PMA (56). These latter findings have lead to the proposal that arachidonate activates at a point distal to other activators. However, McPhail et a2. (57) showed that arachidonate added to detergent-solubilized neutrophil fractions activates protein kinase C. Our results show that activation by arachidonate is blocked at similar concentrations of sphinganine to other activators tested, supporting the involvement of protein kinase C in this activation mechanism.
In the present studies caIcium has not been implicated as a direct activator of the oxidative burst. Data relating to this ion are contradictory, with two calcium ionophores giving different results. lonomycin apparently does not activate (58) while A23187 does (59). This apparent discrepancy may be explained by the finding that A23187 appears to activate the hydrolysis of phosphatidylinositols (48), and may therefore also function by diacylglycerol/protein kinase C-dependent mechanisms. In preliminary studies, we find that activation by A23187 is also blocked by sphinganine, thus supporting this interpretation. Further studies are in progress to ascertain the mechanism of activation by this ionophore.
Summry-h summary, we have shown that sphinganine is a potent inhibitor of the activation of the neutrophil oxidative burst. We have further shown through phosphorylation and competitive binding studies that the site of inhibition by sphinganine appears to be protein kinase C. From the studies utilizing various activators, we propose that the activation mechanisms for not only diacylglycerol and phorbol esters, but also FMLP, arachidonate, and opsonized zymosan converge t o act through protein kinase C. The latter may act directly on the NADPH oxidase (60), or may phosphorylate another regulatory component, such as the cytosolic factor recently reported by Curnutte (61).