Platelet activating factor modulates signal transduction in Dictyostelium

Platelet activating factor (PAF) is a potent mediator of cellular functions, with a wide range of physiological and pathological activities (reviewed by Hanahan, 1986; Braquet et al., 1987; Snyder, 1989; Hwang, 1990). In terms of biological activities, PAF has been involved in, among other processes, aggregation and degranulation of platelets (Benveniste et al., 1979), chemotaxis, adhesiveness and aggregation of leukocytes (O’Flaherty et al., 1981), phagocytosis by human macrophages (Bussolino et al., 1991), smooth muscle contraction and possibly regulation of neural function (Kornecki and Ehrlich, 1988). In clinical research, PAF seems to be a key mediator of several pathological processes, such as inflammation, anaphylaxis, bronchial asthma, graft rejection, ovoimplantation and certain disorders of the immune and central nervous system (Hanahan, 1986; Braquet et al., 1987; Barnes et al., 1989; Snyder, 1989; Saito and Hanahan, 1989). At the cellular level, PAF exerts multiple effects, such as guanylate cyclase activation, adenylate cyclase inhibition, activation of phosphoinositol lipid turnover, mobilization and uptake of calcium, and membrane protein phosphorylation. These effects are apparently mediated by a cell surface receptor which, upon binding of extracellular PAF, interacts with an inhibitory G protein (for reviews see Braquet et al., 1987; Snyder, 1989).

PAF is not restricted to mammalian cells but has also been found to be produced during development of the lower eukaryote Dictyostelium discoideum (Bussolino et al., 1991).

D. discoideum cells grow as free-living amoebae, by feeding on bacteria. Development is triggered by starvation, and results in the amoebae collecting into multicellular mounds, which undergo a series of morphogenetic changes, while differentiating into at least two cell types (Bonner, 1967; Loomis, 1975; Bozzaro, 1992). A major morphoregulatory role in this developmental process is played by cyclic AMP, which is secreted by the cells a few hours after starvation, and acts as a chemoattractant between cells and enhancer of developmentally regulated genes (Konijin et al., 1967; Gerisch et al., 1975; Darmon et al., 1975; Gomer et al., 1986; Chisholm et al., 1984). Both activities are mediated by cAMP binding to membrane receptors (Wurster and Bumann, 1981; Bozzaro et al., 1987; Klein et al., 1988; Saxe et al., 1991), which stimulates, via G proteins (Kumagai et al., 1989) and Ins3P turnover (Europe-Finner and Newell, 1987; Van Haastert et al., 1989), adenylate and guanylate cyclases (Gerisch, 1987; Europe-Finner and Newell, 1985; Janssens and Van Haastert, 1987), as well as mobilization of Ca²⁺ from external and internal pools (Bumann et al., 1986; Europe-Finner and Newell, 1986). Adenylate cyclase activation is necessary for establishment of the cAMP relay (Roos et al., 1977; Gerisch, 1987), whereas Ca²⁺ and guanylate cyclase activation are involved in oriented cell motility (Ross and Newell, 1981; Liu and Newell, 1988).

In our previous study with Dictyostelium cells, the intracellular concentration of PAF was found to increase during development and to undergo oscillatory changes upon stimulation of the cells with cAMP (Bussolino et al., 1991). In contrast to most mammalian cells, where PAF is, at least in part, secreted, in Dictyostelium PAF was undetected extracellularly (Bussolino et al., 1991). To further study the relationship between PAF and cAMP, and the potential involvement of PAF in signal transduction we have investigated in D. discoideum cell suspensions the effects of PAF on the levels of cAMP, cGMP and Ca²⁺. The early steps of D. discoideum development, i.e. from solitary growth-phase cells to multicellular aggregates, occur in agitated cell suspensions. In such suspensions spontaneous periodic activities as well as responses to effectors like cAMP can be continously monitored by means of an optical technique measuring light-scattering (Gerisch and Hess, 1974). Spontaneous periodic light-scattering changes accompany the development of D. discoideum cells to aggregation. Depending on the profile, spike-shaped and sinusoidal oscillations have been distinguished, the first occurring between 3 and 6 h of development and being coordinated by cAMP signals (Gerisch and Wick, 1975). Exogeneous cAMP elicits in these cells two subsequent decreases in light-scattering (Gerisch and Hess, 1974). Light-scattering changes reflect structural changes and indicate alterations of cell shape or the size of cell aggregates or both (Gerisch and Hess, 1974; Wurster and Kurzenberger, 1989).

The attraction of the light-scattering technique is that if PAF induces responses resulting in structural changes of cells or cell aggregates or, likewise, if PAF interferes with spontaneous or cAMP-induced activities it will be immediately observed and the underlying mechanism can subsequently be further analyzed. Our results show that PAF amplifies cAMP-dependent responses and, in addition, stimulates the uptake of Ca²⁺ by the cells.

C₁₈ or C₁₆ PAF (1-O-octadecyl- or 1-O-hexadecyl-2-O-acetyl-sn-glycero-3-phosphocholine) and lyso-PAF (1-O-hexadecyl-sn-glycero-3-phosphocholine) were purchased from Bachem (Bubendorf, Switzerland), cyclic AMP from Boehringer (Mannheim, FRG). The enantiomer S-form of PAF (1-O-octadecyl-2-(S)-acetyl-glycero-3-phosphocholine) was a gift from Drs. H. P. Kertscher and S. Ostermann (Medical Academy, Erfurt, FRG). cAMP and cGMP radioimmunoassay kits were obtained from Amersham (Amersham, England).

D. discoideum AX2 cells were cultured axenically as previously described (Watts and Ashworth, 1970). The amoebae were harvested during exponential growth, washed twice with cold 17 mM Soerensen phosphate buffer, pH 6.0, adjusted to 2 × 10⁷ cells per ml in this buffer and shaken at 150 revs/min until used. Cultivation of cells and all experiments were done at 23°C. In experiments with PAF or lyso-PAF, the lipid molecules were solubilized in 95% ethanol at concentrations high enough to allow a thousand-fold dilution of the solvent upon addition to the cell suspension.

Light-scattering properties of cell suspensions were measured either as described by Gerisch and Hess (1974) or with a modified technique accommodating larger suspension volumes (Wurster and Kurzenberger, 1989).

Ca²⁺ uptake was measured in a cell suspension by using a calcium-ion-sensitive electrode as described by Bumann et al. (1984).

Samples of 2 × 10⁶ cells in 0.1 ml were quenched with 1 vol. of 2 M perchloric acid. After centrifugation, neutralization with potassium carbonate and acetylation, the concentrations of cAMP and cGMP in the extracts were measured using Amersham radioimmunoassay kits.

Since in previous experiments PAF synthesis was found to increase transiently in response to external cAMP (Bussolino et al., 1991) we have now tested whether external PAF affects the period and/or the amplitude of autonomous oscillations. PAF added between two light-scattering spikes caused the subsequent spike to be amplified (Fig. 1A). The enhancing effect did not depend on the time of addition and was restricted to the next spike or sometimes to the next two spikes. The subsequent spikes were often reduced in amplitude. No phase shift or relevant changes in the period of the spikes were observed. The optimal concentration for enhancing the light-scattering changes was 10 μM PAF, although a clear effect was detected also with 1 μM; 0.1 μM was ineffective. The enhancing effect was specific for the naturally occurring PAF (R-enantiomer); the S-enantiomer and lyso-PAF were ineffective (not shown).

Sinusoidal oscillations were not significantly affected by PAF (Fig. 1B). In these experiments the oscillation amplitude was not constant but smaller and bigger waves alternated independently of PAF addition.

Spike-shaped oscillations are accompanied by periodic increases of the cAMP level (Gerisch and Wick, 1975). The finding that PAF enhanced the amplitude of the light-scattering spike, prompted us to examine whether this reflected an increased accumulation of cAMP. Indeed, concomitantly with amplification of an optical spike in response to 10 μM PAF, the cAMP concentration (extracellular + intracellular) was found to increase 2-to 3-fold (Fig. 2).

To analyze the effects of PAF on cAMP-mediated cellular responses more directly, we used cells capable of producing spike-shaped oscillations and entrained them by applying cAMP pulses at intervals of 5 min. Cyclic AMP induces two successive decreases in light-scattering (Gerisch and Hess, 1974). PAF itself failed to induce a similar light-scattering response. However, if added before a cAMP pulse, it amplified the cAMP-induced light-scattering peaks, in particular the second one (Fig. 3).

The cAMP-induced light-scattering changes are accompanied by a fast increase in cellular cGMP followed by a slow increase in both cGMP and cAMP levels, the latter being known as the cAMP relay (Roos et al., 1977; Wurster et al., 1977). As shown in Fig. 4, 2 μM PAF enhanced the cAMP-induced cGMP and cAMP accumulation. Both cGMP peaks and the cAMP peak were amplified by a factor of about 3. The stimulatory effect of PAF on both cGMP peaks is noteworthy. Whereas the enhanced increase in the second peak could be explained as a response to enhanced accumulation of cAMP, enhancement of the first peak occurs in response to an external cAMP stimulus of the same intensity (20 nM) as in the control. Thus PAF must act on a transduction event leading to activation of guanylate cyclase in addition to that of adenylate cyclase.

Spontaneous and induced processes in D. discoideum cells are also accompanied by transient changes of extracellular and intracellular Ca²⁺ levels (Bumann et al., 1986; Europe Finner and Newell, 1986). We examined the potential effects of PAF on Ca²⁺ uptake by continuously monitoring the extracellular Ca²⁺ concentration. Results obtained show that PAF, at concentrations between 0.1 and 10 μM, transiently stimulated Ca²⁺ uptake. After 5 h of development in suspension 1 μM PAF elicited an uptake of Ca comparable in time course and quantity to that of 0.1 μM cAMP (Fig. 5). At earlier times PAF was less effective than cAMP. The stimulation of Ca²⁺ uptake by PAF was direct and did not require the concomitant addition of cAMP. Again, only the naturally occurring R-enantiomer of PAF elicited this reaction, the S-enantiomer and lyso-PAF being ineffective.

The PAF-induced Ca²⁺ uptake was not accompanied by measurable changes in cGMP and cAMP levels (not shown). These results suggest that the effect of PAF on Ca²⁺ influx is not mediated by these nucleotides.

The finding that Dictyostelium cells produce PAF in a developmentally regulated way, and in response to cAMP, has led to the suggestion that PAF could be involved in cAMP signal transduction (Bussolino et al., 1991). In an attempt to test this hypothesis, we have applied to cell suspensions techniques that allow continuous monitoring of activities related to signal transduction. By measuring lightscattering we have shown that PAF interferes with spikeshaped oscillations and cAMP-induced changes, amplifying both responses. In contrast, sinusoidal oscillations, which supersede spike-shaped oscillations at about six hours of development, are unaffected.

Spike-shaped oscillations are regulated by oscillatory changes in the concentration level of extracellular cAMP, which is secreted by the cells (Gerisch and Wick, 1975). As expected from the amplifying effect of PAF on lightscattering responses, the level of total (intracellular and extracellular) cAMP was found to be significantly increased by PAF. It must be emphasized, however, that in the absence of spontaneous or cAMP-induced cAMP relay, PAF does not stimulate any change in cAMP concentration. The same holds true for the intracellular accumulation of cGMP, which appears to be stimulated by PAF, but only in conjunction with exogenously added cAMP.

In a second set of experiments the effect of PAF on the extracellular Ca²⁺ concentration was examined. In contrast to the indirect effect of PAF on cAMP and cGMP changes, PAF directly induced a transient decrease in the extracellular Ca ²⁺ concentration.

Only the naturally occurring R-enantiomer of PAF shows an effect on light-scattering and calcium responses, indicating that PAF stimulation probably results from interaction with specific target sites. These interactions could be mediated by a specific cell surface receptor or by an intracellular acceptor protein. Both cell surface and intracellular PAF receptors have been described in several cell types (Hwang, 1990). Experiments aimed to analyze binding of radiolabelled PAF to living cells failed to detect specific cell surface receptors (Sordano et al., unpublished). The present results, therefore, imply that in Dictyostelium a specific intracellular PAF acceptor exists. In agreement with this assumption are experiments showing that PAF is rapidly taken up and metabolized by Dictyostelium cells (Sordano et al., unpublished). The absence of specific membrane receptors and the rapid degradation of PAF could explain the relatively high concentrations of PAF which are required for the effects described in this paper.

The effects of PAF on signal transduction reactions in Dictyostelium are in some aspects similar to those observed in mammalian cells but in other aspects show interesting differences. In Dictyostelium, like in mammalian cells (Schwertschlag and Whorton, 1988; Snyder, 1989), transient cellular responses to PAF are reduced or not evoked following a second exposure of the cells to PAF. Whether this is due to desensitization, as apparently is the case for mammalian cells, remains to be established. Both in mammalian (Lee et al., 1981) and in Dictyostelium cells, PAF elicits an influx of Ca²⁺. However, unlike the results obtained with Dictyostelium, in mammalian cells PAF causes inhibition of adenylate cyclase. The latter appears to result from the activation by PAF of an inhibitory G protein (Avdonin et al., 1985).

G proteins also exist in Dictyostelium and mediate the cAMP-induced activation of adenylate cyclase, guanylate cyclase and the inositolphosphate cycle (Kimmel and Firtel, 1991). To explain in this molecular framework the PAF effects described here we assume that PAF acts on a transducer protein (intracellular acceptor), which in turn opens the membrane Ca²⁺ channel and also transforms a G protein into a form which upon stimulation by cyclic AMP is a more potent activator of the pathways leading to cAMP and cGMP. Alternatively to the G protein, PAF could act at the level of the cAMP receptor, by modulating its isoforms (Janssens and Van Haastert, 1987), and thus influencing either cAMP binding or desensitization of the receptor(s). Further experiments should distinguish between these hypotheses. Although in mammalian cells PAF is secreted and acts by binding to membrane receptors, an intracellular role for PAF has been postulated, based on the observation that most of the produced PAF remains intracellular and that intracellular acceptors for PAF appear to exist (Stewart et al., 1990; Hwang, 1990). In Dictyostelium, PAF is not detected in the extracellular medium (Bussolino et al., 1991; Wurster, Bussolino and Bozzaro, unpublished results). Therefore, Dictyostelium appears to be a suitable experimental model for studying the intracellular role of PAF.

This work was supported by a NATO Collaborative Research Grant (no. 900992).