Prostaglandin cyclooxygenase products but not thromboxane A2 are involved in the pathogenesis of ewthromelalgia in thrombocythaemia

Fluid of artificial blisters from erythromelalgic skin areas in primary thrombocythaemia contained a high amount of prostaglandin-E-like activity. Dazoxiben did not alleviate the erythromelalgia in patients with primary thrombocythaemia despite complete inhibition of platelet malondialdehyde and thromboxane B2 synthesis and no inhibition of prostaglandin-E-like material. During a 10-day dazoxiben treatment period, persistent erythromelalgia was associated with a significant shortened mean platelet life span of 3.2 days. During subsequent treatment with low dose acetylsalicylic acid daily complete relief of erythromelalgia was associated with inhibition of platelet prostaglandin endoperoxide production and correction of platelet mean life span to normal, 7.9 days. These observations indicate that prostaglandin E2, or another prostaglandin endoperoxide metabolite, is involved in the pathogenesisof erythromelalgia. The presented study does not give one single clue as to the origin (platelet, vessel wall or other) of the prostanoid, but very likely originates from platelets because a very low dose of acetylsalicylic acid (250 to 500 mg every other day), which irreversibly inhibits platelet cyclooxygenase, is highly effective in the prevention of erythromelalgia in thrombocythaemia.


Introduction
The alveolar macrophage (AM) is the most abundant cell in the airway lumen and may be an important contributor to the inflammation associated with pulmonary diseases such as asthma, a'2 Isolated AM respond to a variety of immunological and non-immunological stimuli by generating both acute-phase and pro-inflammatory mediators. The acute-phase mediators include the spasmogens thromboxane A2, leukotriene C4 and platelet activating factor (PAF) as well as reactive oxygen species and lysosomal enzymes which can cause local tissue damage. 3 Among the pro-inflammatory mediators are PAF, leukotriene B 4 and a number of cytokines which are chemoattractant to mast cells, eosinophils and lymphocytes, all of which have been implicated in pulmonary inflammation, a'2 In addition to being directly activated AM can be primed in vitro by agents including LPS 4 and gamma interferon to give an exaggerated response to subsequent stimuli. AM primed in vitro may be comparable with AM from asthmatic subjects which have been shown to release greater amounts of many of the above mediators on stimulation than cells from non-asthmatic subjects. [6][7][8][9] The biochemical mechanisms underlying priming are poorly understood but may involve amplification of signal transduction processes. For the chemotactic peptide FMLP three receptor-coupled 993  were killed by intraperitoneal administration of sodium pentobarbitone, the trachea cannulated and bronchoalveolar lavage carried out with two 5 ml volumes of saline at 37C. Lavage fluid was centrifuged at 300 x g" for 8 min, cells resuspended in Ca 2+ and Mg 2+ free HBSS and pooled. The cell suspension was centrifuged again and cells resuspended at 106 cells/ml in RPMI 1640 medium supplemented with 1% foetal calf serum, 50 U/ml penicillin and 50 #g/ml streptomycin. For superoxide experiments cells were plated out in 24-well dishes at 106 cells/well; for TXB2 experiments in 96-well dishes at 2 x 105 cells/well. After 2 h in an incubator (37C, 5% COg) non-adherent cells and debris were removed by changing the medium. Adherent AM were left in the incubator for 18-20 h before use. Measurement of superoxide production: Tissue culture medium was removed, AM washed once with HBSS then HBSS (0.5ml) with or without inhibitors was added and the cells incubated at 37C for 2 h. Buffer was then replaced by fresh HBSS (0.5 ml) containing Fe 3+ cytochrome c (55/iM), again with or without inhibitors. After 15 min at 37C FMLP or PMA was added and the incubation continued for a further 30 min. Superoxide generation was determined from the difference in absorbance of the AM supernatants at 550 nm in comparison with supernatants from cells stimulated in the presence of superoxide dismutase (0.1 mg/ml). 19 Using this protocol the absorbance difference was directly proportional to superoxide concentration.
Measurement of TXBe release" After removal of tissue culture medium AM were washed once with HBSS then incubated for 2 h at 37C in HBSS (0.2 ml) with or without inhibitors and with or without LPS (10 ng/ml) for cell priming. The buffer was replaced by fresh HBSS (0.2 ml) again with or without inhibitors, and the incubation continued for 15 min prior to adding FMLP for a further 15 min. Supernatants removed at appropriate times were stored at -20C prior to quantification of TXB 2 by radioimmunoassay. Samples were diluted to fall within the assay standard curve; concentrations were calculated by interpolation using a Multi Calc software program (Wallac, Finland). Statistics" Superoxide and TXB 2 concentrations were calculated as mean --t-_ S.E.M. of triplicate or quadruplicate determinations within each experiment. Since the magnitude of stimulated release varied with different preparations of cells, in all cases the figures show values from one representative experiment which was conducted 2-4 times. 0.10 _---t-0.01 to 1.0 +_ 0.2ng/10/g AM protein (n 4 experiments). However LPS potentiated markedly (primed) the response to FMLP over the full concentration range tested (0.3-30 nM) (Fig. 2). FMLP (3 nM), as a non-maximal stimulus, was used to study the eEects of Ro 31-8220 and genistein on TXB 2 release. When present during the 2 h pre-incubation, or 2h LPS priming, stage and the subsequent challenge stage genistein (1-30/M) dose-dependently inhibited FMLP stimulated TXB2 release in both non-primed and primed AM with similar ICs0 values (17 and 18#M, respectively, in the experiment shown in Fig. 3A). Under these conditions Ro 31-8220 (0.1-10/,M) did not inhibit FMLP induced TXB 2 release in non-primed AM but tended rather to potentiate release at higher concentrations (by >27% at 3 and 10 #M in the experiment shown) (Fig. 3B). The extent of potentiation varied between experiments but was a consistent observation. In LPS primed AM low concentrations of Ro 31-8220 (0.1-3 #M) had no effect on the FMLP response whereas Ro 31-8220 at 10/,M inhibited TXB. release by 80% relative to the LPS baseline level. Since this suggested that Ro 31-8220 had an inhibitory effect on a component of LPS priming, further studies were performed in which Ro 31-8220 (10 M) was added to the AM only during selected stages of priming and/or activation (Fig. 4). Ro 31-8220 had no effect on the LPS primed response when present only during FMLP challenge, inhibited the response by approx. 50% when present only during priming, but had greatest effect, reducing FMLP induced TXB2 release by 80 and 100% in relation to the LPS baseline value (n--2 experiments), when present during both priming and challenge. The essential requirement for inhibition was thus that Ro 31-8220 was present during the LPS priming stage.

Discussion
Ro 31-8220 and genistein were used in the present study as selective inhibitors of PKC and PTK, respectively. Ro 31-8220 is reported to have a 100-fold selectivity for PKC over protein kinase A (PKA) and a 1 000-fold selectivity over Ca2+-cal modulin dependent kinase 18 and has been used by several groups as a selective PKC inhibitor in intact cells.  Genistein inhibits a number of PTKs but shows very weak activity against serine and threonine kinases including PKC, PKA and phosphorylase kinase. 17 23 and zymosan induced phospholipase C activity in liver macrophages. 1 In addition, in the latter study the activation of phospholipase C by zymosan was enhanced by chronic pretreatment of the macrophages with PMA to down-regulate PKC. It is unlikely that a Ro 31-8220 insensitive isoform of PKC was involved in FMLP stimulated superoxide and TXB 2 production in the AM studied in the present work since Ro 31-8220 is not isozyme selective and, in addition, the compound inhibited the superoxide response to PMA. In contrast to the complex activity of Ro 31-8220, the consistent inhibitory effects of genistein on FMLP responses in both non-primed and primed AM support only a proactive role for a PTK in FMLP stimulus-response coupling in AM, as has recently been reported for neutrophils 2'24 and for HL-60 cells. 13 Whether the pathways leading to superoxide or TXB 2 release are separately regulated by PTK or whether PTK action affects an early common event such as PLCy activation, 4 or PLD activation, is each of which has been demonstrated in other cell types, remains to be determined.
Nonetheless, the ability of genistein to inhibit superoxide production and TXB 2 release almost completely emphasizes the key importance of PTK activation following FMLP receptor stimulation.
The observation that Ro 31-8220 was able to inhibit FMLP induced TXB 2 production in primed AM provided that the compound was present during the priming event is of interest in demonstrating that LPS induces priming by a process which is positively regulated by PKC. This does not appear surprising in view of other reports that LPS can activate PKC in macrophages 26 and that responses to LPS such as cytokine production are blocked by PKC inhibitors. 27 It does, however, differ from findings in P388 D1 macrophages where the non-selective PKC inhibitor H7 did not affect cell priming. 16 The fact that Ro 31-8220 present during FMLP challenge as well as during priming increased the inhibition of TXB 2 release contrasts with the potentiation of release seen in non-primed AM and probably reflects the complex interplay between feed forward and feed back effects of PKC on mediator release in these cells.
In summary, using selective enzyme inhibitors it was found that activation of PTK is a key event in FMLP generation of superoxide and TXB 2 in guinea-pig AM whereas PKC either is not activated, or is activated but has counter-balancing feed back and feed forward effects on mediator release in non-primed cells. On the other hand activation of PKC appears to be required for LPS priming of these cells for enhanced TXB2 release. Thus both PTK and PKC inhibitors could conceivably be effective in reducing the release of these mediators from macrophages in the airways in vivo under conditions where priming has occurred during an inflammatory response.