Release of adenosine from human neutrophils stimulated by platelet activating factor, leukotriene B4 and opsonized zymosan

Isolated human polymorphonuclear leukocytes (PMNL) stimulated by platelet activating factor (PAF), leukotriene B4 (LTB4) or opsonized zymosan (OZ) released adenosine measured by thermospray high performance liquid chromatography mass spectrometry in the cell-free supernatants. Stimulation by PAF or LTB4 resulted in a bellshaped concentration-effect curve; 5 × 10−7 M PAF, 10−8 M LTB4 and 500 μg ml−1 OZ induced peak adenosine release, thus cytotoxic concentrations did not elevate adenosine level in the supernatants. Therefore adenosine release was characteristic of viable cells. As calculated from concentration-effect curves, the rank order of potency for adenosine release was PAF > LTB > OZ. These resuits suggest that adenosine, when bound specifically to membrane receptor sites, may initiate signal transduction, and, in co-operation with other inflammatory mediators, may modulate phagocyte function, e.g. production of chemoluminescence (CL).


Introduction
Platelet activating factor (PAF) is a phospholipid autacoid implicated as mediator in the pathogenesis of inflammation, thrombosis, immune disorders, septic shock and a great variety of physiological and pathophysiological conditions. 1'2 PAF as second messenger of diverse injurious stimuli releases eicosanoids and superoxide anions from leukocytes, macrophages and endothelial cells. Leukotrienes (LTs) are metabolites of arachidonic acid (AA) formed by 5'-lipoxygenase.
One leukotriene, LTB4, is a potent chemotactic and aggregating agent released from polymorphonuclear leukocytes (PMNL). 4'5 LTB 4 is also involved in a variety of pathophysiological processes, including y-interferon production.
Phagocytosis induced by opsonized zymosan (OZ) is one of the most widely used models for testing the function of stimulated PMNL. Engulfment of particles via Fc, CR1 and CR3 receptors involves marked changes in cellular metabolism, leading to degranulation and production of superoxide anions.
Adenosine is a natural nucleoside known to regulate various cellular functions, including neurotransmission and local circulation. 1 These effects of adenosine appear to be mediated by two separate subtypes of binding sites, i.e., A and a 2 receptors. 1 By interacting with one of these receptor subtypes, adenosine can initiate a transmembrane signal which then may inhibit or (C) 1992 Rapid Communications of Oxford Ltd stimulate adenylate cyclase via A1 or A 2 receptors, respectively. A third adenosine recognition site, termed the P site, located on the catalytic subunit of adenylate cyclase, is activated by relatively high concentrations of adenosine. 2 Near micromolar concentrations of adenosine, interacting with A 2 receptors, have been shown to inhibit PMNL functions. 13 The present experiments were designed to study the production and release of adenosine from PMNL stimulated by PAF, LTB4 and OZ. The results obtained suggest that adenosine released from PMNL may induce inhibitory action, which then, as a regulatory feedback signal, may protect the phagocytes from irreversible damage due to overstimulation during the inflammatory process. The mass spectrometer monitored the eluent continuously at m/z 268. Measurement of chemiluminescence" 2 106 PMNL were incubated in the presence of 5 x 10 -7 M PAF, dissolved in PBS/BSA, and/or with 2-chloroadenosine in various experimental conditions at 37C for 60 min. Chemiluminescence of PMNL was measured in PBS at a final volume of 1 ml in the presence of 10 -7 M luminol using a Nuclear Chicago/300 liquid scintillation counter (Searle, Indianapolis, IN) in the off coincidence mode. 19 '2 Viability of cells" Viability was determined by the trypan-blue exclusion test to detect the percentage of viable cells at the beginning of the experiments and to check the cytotoxic elect of PAF, LTB 4 and OZ at the end of incubation with these stimulating agents. 2 Statistical analysis: Data are expressed as means __+ standard deviation (SD). Each adenosine determi-nation was carried out in triplicate samples. Data were statistically analyzed by the tailed Student's t-test. Differences were considered significant when p was less than 0.05.

Results
Adenosine release: Adenosine release was studied in samples of 2 106 PMNL m1-1 incubated with various concentrations of PAF, LTB 4  by LTB4, in the concentration ranges of 10 -9 to 10 -4 M or 10 -1 to 10 -s M PAF or LTB 4 respectively was 2.25, indicating an efficacy for PAF higher than for LTB 4. The kinetics of adenosine release were determined in PMNL stimulated with 5 x 10 -r M PAF, 10 -8 M LTB 4 and 500 #g m1-1 OZ at 37C for 30 and 60 min. This entirely non-cytotoxic concentration of OZ was used since it was shown to induce the highest CL of PMNL. 22 As shown in Table 1, adenosine concentrations in the supernatants of cell suspensions stimulated for 30 min were lower than that measured after 60min stimulation. Stimulation by OZ for 30 min did not produce any detectable elevation of adenosine concentration. The rank order of potency for adenosine release was PAF > LTB 4 > OZ.
Chemiluminescence: The inhibition by adenosine of CL in PMNL was demonstrated in transfer experiments. Supernatants of PMNL incubated in the presence of 5 10 -7 M PAF, containing approximately 400nM adenosine, were transferred to unstimulated cells, and the luminol induced  amplification of CL in these cells was measured, or PAF induced CL was determined in the presence of 2-chloroadenosine, a stable analogue of adenosine. Preincubation of PMNL with PAF at 37C for 60min produced adenosine concentrations that were able to inhibit the CL of freshly added PAF when transferred to unstimulated cells. This stimulation was almost as intense as that was seen after addition of 10 -6 M 2-chloroadenosine ( Table  2).

Discussion
Beyond the lipid character, common features in the effects of PAF and LTB 4 are the induction of chemotaxis, aggregation and superoxide anion production in PMNL. 1'23'24 This study shows that both PAF and LTB4 release adenosine from PMNL, although the effect of PAF is more marked than that of LTB4. The dit:ference may be explained by the fact that PAF can also release LTB4, 3'18 thus adenosine release induced by PAF presumably includes that released by LTB 4. Regardless of distinct receptor binding sites for PAF and LTB4 on PMNL membrane, <24-26 these findings point to the similarity of signal transduction triggered by the two autacoids, suggesting their involvement in a common pathway of the inflammatory process. The particles of OZ are internalized by PMNL via Fc, CR1 and CR3 receptors leading to degranulation of specific azurophil granules and production of superoxide anion. 27 In our study, OZ proved to be the least effective in releasing adenosine from human PMNL. Stimulated PMNL produce PAF and LTB4 at picomolar concentrations, 28 and this may explain the finding that adenosine concentrations are much lower in supernatants of cells stimulated with OZ than in supernatants stimulated with higher, micromolar concentrations of exogenous PAF or LTB 4. This can also be reflected by the kinetics of adenosine production in stimulated cells. While 30min stimulation was found to be optimal for production of adenosine by PAF or LTB4, this incubation period was insufficient in OZ-activated PMNL 17'29 to detect a measurable amount of adenosine in the supernatant, while 60 min stimulation resulted in a well established increase of adenosine concentration. Accordingly, adenosine production by PAF and LTB 4 was also more marked after 60 min than at 30 min stimulation.
The preparation of a completely platelet free human PMNL suspension is practically impossible. The rate of platelet contamination in our PMNL suspensions was ordinarily 1:1. The estimated amount of adenosine possibly derived from aggregated platelets was 50nM. To aggregate platelets but not neutrophils, ADP was added to cells suspended in TC-199 medium, and adenosine release was measured in the supernatants 60 min later (data not shown). From these results the conclusion can be drawn that the major part of adenosine released by PAF, LTB 4 and OZ is derived from PMNL.
Adenosine is produced by the breakdown of intracellular ATP, and an increased consumption of ATP results from the stimulation of phagocytes via the pathway of ATP synthesis from ADP: 5'-nucleosidase AMP adenosine 22 (2) We assume that, at a certain degree of ATP depletion in activated cells, this process may lead to accumulation of adenosine at nearly micromolar concentrations in the extracellular space, because some adenosine molecules may escape from the rapid breakdown by adenosine deaminase located on the external surface of cells. 3 These molecules may then bring important signal transduction for regulating the function of surrounding cells. Adenosine binding A 2 receptor subtypes and P sites, has been shown to inhibit PMNL functions, e.g. intracellular killing or generation of oxygen derived free radicals. 31 As previously described, 11 this effect is related to an activation of adenylate cyclase with a concomitant elevation of intracellular cAMP level. Intracellular cAMP raised by eicosanoids, in particular prostacyclin (PGI2) has been shown to downregulate eicosanoid production in platelets 32'33 and vascular endothelium. 34 PAF has been shown to release AA by conversion not only to the lipoxygenase product LTB4,3'18 but also through the cyclooxygenase pathway to prostaglandin E 2 (PGE2) 35 and PGI2 .36 All these events have been shown to augment the number of emitted photons in luminol induced CL. 37 Eicosanoid synthesis by itself is therefore associated with increased light emission 38 and indomethacin can block such Ct. 39 Consequently, the elevation of intracellular cAMP by PGs, the endproducts of AA metabolism, is accompanied by decreased CL as a consequence of feedback inhibition brought about either directly as a result of decreased precursor conversion or through the NADPH oxidase system. 4 Thus, CL of phagocytes can be modulated not only by adenosine released from stimulated cells, but also by AA and its biologically highly active endproducts depending upon the stage of their production. At the same time, a great variety of interactions may occur among inflammatory mediators, for example, adenosine can block LT synthesis. 41 An elevated intracellular cAMP level, induced by either adenosine, 42 PGE241 or PAF 43 may be an important signal transduction to downregulate eL, 44 thus being a sensitive marker for the metabolic and functional state of phagocytes. Another important aspect of this autoregulatory feedback network is that cAMP has also been implicated in the inhibition of PAF release. 45 On the other hand, the extracellular nucleotides, such as ATP, ADP and AMP, may induce the opposite effect, i.e., stimulation of adhesiveness and other functions of PMNL. 46 This regulatory and inhibitory role for exogenously applied and endogenously released adenosine has been described in stimulated PMNL. [47][48][49] Our data confirm these observations and support the view that production and release of adenosine may regulate the function of activated PMNL and other phagocytes. In the autocatalytic feedback network of inflammatory mediator release, s adenosine may therefore be regarded as an important signal molecule downregulating the production of PAF, LTB 4 and other lipid mediators, and through this mechanism activated PMNL and other phagocytes may be protected from potentially irreversible damage due to overstimulation, s1'52 This modulatory action, however, takes place in co-operation with other mediators, mainly AA derivatives.