Phenylglyoxal is not a selective inhibitor of phagocytosis

Polymorphonuclear leukocytes (PMN) move to, bind and ingest foreign material in vivo and in vitro. Binding of particles results in stimulation of the oxidative metabolism and release of lysosomal enzymes from these cells. To investigate the effect of binding to, and/or uptake of particles by, PMN on the initiation of several processes in these cells, a well defined agent is needed that specifically inhibits phagocytosis without affecting other cell functions. In the past, metabolic inhibitors such as N-ethyl maleimide, fluoride, or mono-iodoacetic acid have been used to inhibit ingestion. Each of these (non-specific) agents exerts numerous side effects, however (Curnutte & Babior, 1975; Mandell, 1972; Noseworthy & Karnovsky, 1972). In 1971 the fungal metabolite cytochalasin B was introduced for this purpose (Davis, Estensen & Quie, 1971). Presumably, this compound acts by disrupting microfilaments (Allison & Davies, 1974; Axline & Reaven, 1974; Carter, 1967; Spooner & Wessels, 1970) and, as a result, not only phagocytosis but also other cell functions are affected (Carter, 1967; Estensen & Plagemann, 1972; Kletzien, Perdue & Springer, 1972; Plagemann & Estensen, 1972; Roos, Homan-Muller & Weening, 1976; Zigmond & Hirsch, 1972).

Recently, phenylglyoxal (C₆H₅COCHO), an agent which reacts with guanido groups of arginine residues (Takahashi, 1968) has been reported selectively to inhibit endocytosis in phagocytes (Oliver, Yin & Berlin, 1976; Yin & Lu, 1975). Phagocytosis of oil droplets by glycogen-elicited rabbit PMN and of polyvinyl toluene particles by resident mouse peritoneal macrophages is inhibited by phenylglyoxal, as is fluid phase pinocytosis, cell spreading, and - at higher phenylglyoxal concentrations - formation of Concanavalin A caps (Oliver et al. 1976; Yin, 1975). This inhibition appears to be reversible. Phenylglyoxal shows no effect on other cell properties, such as adenine transport, lipid fluidity or adhesion. Moreover, neither the ATP level nor the cell shape is changed by this agent, and no cross-linking between membrane proteins seems to be introduced. However, because the membrane-associated release of lysosomal enzymes is also inhibited by phenylglyoxal (Yin, 1975), it must be concluded that this drug affects endo-as well as exocytic processes.

To ascertain whether phenylglyoxal is a useful agent for the investigation of the relation between stimulus binding and phagocyte activation, we have tested this compound on several other phagocyte functions. The results indicate that phenyl glyoxal is not a specific inhibitor of endo- or exocytosis.

(K & K Laboratories, New York, U.S.A.) was dissolved in 10 mM phosphate buffer (pH rz) at a concentration of 1 mg/ml. A fresh solution was made each day.

(LC.I. Research Laboratories, Alderley Park, Cheshire, UK) was prepared as a 50 µg/ml stock solution in 1 % dimethyl sulphoxide.

(Consolidated Midland Corp., Brewster, N.Y., U.S.A.) was prepared as a 200 µg/ml stock solution in 10 % dimethyl sulphoxide.

Zymosan was opsonized with fresh normal human serum as described previously (Goldstein, Roos, Kaplan & Weissmann, 1975). A final concentration of r mg/ml was used in the cell incubations. Specific opsonization of zymosan with either immuno globulin G or complement component C3 was performed as described before (Roos, de Boer & Weening, 1977a).

¹²⁵1-labelled STZ. To obtain ¹²⁵1-labelled STZ, zymosan particles were iodinated with the lactoperoxidase method, as follows. Zymosan (18 mg), lactoperoxidase (3 U.) and Na¹²⁵I (0·05 nmol, 2 mCi) were incubated for 9 min at room temperature in a volume of 5 ml. At o, 3 and 6 min, 0 · 5 µ mol H₂O₂ was added. The particles were spun down (4500 g, 5 min, room temperature) and washed with 154 mM NaCl until the radioactivity of the medium was less than twice the background (8-10 washings). The labelling efficiency was 50-60 %. The ¹²⁵1-labelled zymosan was opsonized with serum as described above. Before use, the ¹²⁵1-STZ was centrifuged over Ficoll-Isopaque (d = I · 095 g/cm³) at 2000 g and room temperature for 10 min to remove light particles. The ¹²⁰1-STZ was washed twice and resuspended in 154 mM NaCl.

Ficoll-Isopaque of 1 ·095 g/cm³ was prepared by adding 169 · 3 g Ficoll (Pharmacia, Uppsala, Sweden) to 76 · 6 ml 1 · 156 M metrizoate (Isopaque) (Nyegaard, Oslo, Norway) in 122 · 8 ml 0 · 175 M Tris-HCI (pH 7 · 4 at o °C), and the final volume was made up to 1000 ml with distilled water (Loos & Roos, 1974).

Polymorphonuclear leukocytes were isolated from fresh defibrinated human blood as described elsewhere (Weening, Roos & Loos, 1974), and suspended in a buffer con sisting of 138 mM NaCl, 2·7 mM KCI, 8·1 mM Na₂HPO₄, 1·5 mM KH₂PO₄, o·6 mM CaCl₂, 1·0 mM MgCl₂, 5·5 mM glucose and 0·5 % (w/v) human albumin, pH 7·4. The cells were counted electronically (Coulter counter, model ZF). The final cell suspension contained more than 95 % granulocytcs, as judged microscopically after nuclear staining.

Chemotactic activity towards casein (1 mg/ml) was measured in modified Boyden chambers with the’ leading front’ method of Zigmond & Hirsch (1972). The distance of cell penetration into Millipore filters with 3- µ m pore size (cat. no. SSWP 02500) was measured in µ m. The cell penetration into these filters in the absence of casein was taken as the spontaneous mobility.

Adherence of STZ to PMN was measured by rosette formation as de scribed by Wong & Wilson (1975).

Uptake (plus adherence) of STZ was measured with ¹²⁵1-labelled STZ (Reiss & Roos, 1978). After incubation of 4 x 10⁶cells at 37 °C with 1 mg ¹²⁵1-STZ/ml, ingestion was stopped with NaF (final concentration 10 mM). The cells were separated from non phagocytosed particles by centrifugation over Ficoll-Isopaque of 1 · 095 g/ cm³ (3000 g, room temperature, 10 min). The cells, together with adherent and phagocytosed particles, remained in the interface, whereas the non-phagocytoscd particles were spun down. The radioactivities in the pellet and in the remainder were counted separately, and the number of adherent and phago cytosed particles was expressed as the percentage of the total amount of radioactivity in each sample.

Uptake of ¹⁴C-S. aureus (strain Oxford) was measured with a modification of the method described by Tan, Watanakunakorn & Phair (1971). The bacteria were cultured overnight at 37 °C with 5 µCi U-¹⁴C-protein hydrolysate (56 mCi/ mAtom Carbon, The Radiochemical Centre, Amersham, UK; Code CFB 25), washed and suspended in cell medium to a concentration of about 3 x 10⁸ /ml. PMN were incubated at 37 °C with S. aureus in a ratio of about 1: 3 with 10% (vol./vol.) fresh human serum in a total volume of 5 ml. To stop the uptake of bacteria, samples of 0 · 5 ml of the incubation mixture were added to 1 ml ice-cold saline (140 mM NaCl) buffered with 5 · 8 mM phosphate (pH 7 · 2) and supplemented with 10 % human albumin and 20 mM NaF. Non-ingested bacteria were lysed by a 15-min incubation at 37 °C with 0 · 3 U. lysostaphin (Schwartz/Mann, Orangeburg, N.Y., U.S.A.). The cells were spun down (600 g, o °C, 10 min) washed twice with ice-cold NaF containing medium and dissolved in 1 · 25 ml Scluene-100 (Packard-Becker, Groningen,The Netherlands). After addition of 10 ml scintillation fluid (PPO-POPOP-toluene), the radio activity was measured in a liquid scintillation counter. The variance of this test is 4 % (calcula ted from 100 duplicate values).

Cell respiration was measured polarographically with an oxygen elec trode (Yellow Springs Instrument Co., Yellow Springs, Ohio, USA, model 5331), as described elsewhere (Weening et al. 1974). Phenylglyoxal (200 µ g/ml) had no effect on the O₂ consump tion of glucose+ glucose oxidase.

The production of hydrogen peroxide was measured in cell free supernatants in the presence of 2 mM NaN₃ by a method based on the oxidation of leuko diacetyl-2,7-dichlorofluorescein to a fluorescent compound by H₂O₂ (Homan-Mi.iller, \Veening & Roos, 1975). Phenylglyoxal (200 µ g/ml) decreased the measurements of hydrogen peroxide standards by 35 %. The values found in the samples with phenylglyoxal were corrected for this effect.

Release of lysozyme and /J-glucuronidase was measured as described previously (Goldstein et al. 1975). Extracellular lactate dehydrogenase was measured as a control for cell integrity. Less than 5 % of the total activity of this enzyme in the cells was released under the conditions used in these experiments. Phenylglyoxal (200 µ g/ml) had no effect on the measurements of these enzymes.

Cell viability was assessed with fluorescein diacetate hydrolysis, as described in the preliminary operation manual of the Cytograf, Bio/Physics Systems (Operation Manual, 1974).

The ATP content of the cells was measured with the firefly method (Stanley & Williams, 1969). Phenylglyoxal (200 µg/ml) had no effect on the measurement of ATP standards.

The level of reduced glutathione (GSH) in PMN was measured with a modification (Voetman, Loos & Roos, in preparation) of the method described by Koivusalo & Uotila (1974), in which reduced glutathione participates as a rate-limiting co-factor in the formaldehyde dehydrogenase reaction. In short, cells were lysed with I mM EDTA and introduced into an AutoAnalyzer system with NAO, formaldehyde and formaldehyde dehydrogenase. The production of NADH was measured fluorimetrically. Phenylglyoxal (200 µg/ml) had no effect on the measurement of GSH standards.

Fig. 1. shows a dose-dependent inhibition of phenylglyoxal on the binding plus uptake of STZ. This process was inhibited by more than 90% when PMN were pretreated for 30 min with 100 µg phenylglyoxal/ml at 37 °C. Both the initial rate of phagocytosis and the total uptake of STZ were inhibited. Resuspension of phenylglyoxal-treated cells in medium without phenylglyoxal restored the phagocytic capacity to about 60 %- Similar results were obtained when phenylglyoxal was tested on the phagocytosis of Staphylococcus aureus.

Next, the separate effect of phenylglyoxal on adherence was measured. Table 1 shows that phenylglyoxal induced a significant decrease in the percentage of rosettes with STZ, as compared with cytochalasin B-treated cells. Resuspension of phenyl glyoxal-treated PMN in medium without phenylglyoxal largely restored rosette formation.

Table 2 shows that phenylglyoxal caused a dose-dependent inhibition of the chemotactic response of the PMN towards casein. This inhibition was partially abolished after resuspension of the cells in medium without phenylglyoxal. The spontaneous mobility of the cells was also inhibited by phenylglyoxal and was com pletely restored by resuspension. Phenylglyoxal (20-200 µg/ml) had no effect on the oxygen consumption of resting PMN. Preincubation of PMN with phenylglyoxal resulted in a dose-dependent inhibition of the STZ-stimulated respiration (Fig. 2). At low concentrations of phenylglyoxal, resuspension of the cells in a medium without phenylglyoxal partially restored this oxygen consumption. At concentrations of phenylglyoxal > 200 µ g/ml, however, the inhibition was not released by resuspension. When phenylglyoxal was added after 7 min to non-treated, phagocytosing PMN, the oxygen consumption always decreased within 2 min to about 10% of the original value. This phenomenon was observed with PMN stimulated by Staphylococcus aureus as well as with PMN stimulated by zymosan opsonized either with IgG, C 3, or both.

Similar results were obtained with the soluble stimulator phorbolmyristole-acetate (PMA). Preincubation (30 min at 37 °C) of PMN with phenylglyoxal (100 µg/ml) resulted in a loss of respiratory stimulation by PMA (500 µg/ml). When phenyl glyoxal (100 µg/ml) was added to non-treated PMA-stimulated PMN, the oxygen consumption was completely inhibited within a few minutes.

The hydrogen peroxide production was also affected by phenylglyoxal. Preincubation of PMN with phenylglyoxal (10-200 µg/ml) had no effect on the H₂O₂ production by resting PMN, but resulted in a dose-dependent inhibition of the STZ-stimulated reaction (without additions, 135 nmol H₂O₂/ 10⁶ cells in 30 min; with 10 µg phenyl glyoxal/ml, 85 % inhibition; with 20 µg/ml, 90 % inhibition; with higher concen trations, about 95 % inhibition).

Addition of phenylglyoxal (100 µg/ml) to resting PMN did not cause release of lysosomal enzymes or lactate dehydrogenase. Fig. 3 shows that pretreatment of PMN with phenylglyoxal resulted in a dose-dependent inhibition of lysozyme release during stimulation with STZ. Resuspension of phenylglyoxal-treated PMN in medium without phenylglyoxal had hardly any effect on this inhibition, as shown in Fig. 4. Similar results were obtained with PMA-stimulated cells. None of these treatments caused release of lactate dehydrogenase.

The ATP content of the PMN (1·0±0·4nmol/10⁶ cells, mean±s.o., n = 5) was not significantly affected by phenylglyoxal during 30 min incubation at 37 °C. Moreover, cell viability, as judged by hydrolysis of intracellular fluorescein diacetate, was not changed by phenylglyoxal treatment.

Because phenylglyoxal has been reported to react with reduced glutathione in viiro (Schubert, 1935), the effect of phenylglyoxal on the level of reduced glutathione in PMN was investigated. Addition of 200 µg phenylglyoxal/ml to PMN immediately decreased the intracellular GSH level from 13·5 to 8·4 nmol/10⁷ PMN (mean of 2 experiments). Incubation at 37 °C did not further affect the GSH level in the cells. Phenylglyoxal (200 µg/ml) added to sonicated PMN decreased GSH from 13·5 nmol/ 10⁷ PMN to zero in 30 min at 37 °C. Phenylglyoxal reacted neither with GSH in cell-free medium nor with GSH in 1 mM EDTA-lysed cell samples. This indicates that phenylglyoxal reacted with GSH only in the presence of a cation-sensitive cofactor. It also excludes reaction of phenylglyoxal with GSH after a sample from the incubations was diluted with EDTA.

Finally, since phenylglyoxal effectively stopped the oxygen consumption of PMN even after stimulation with STZ had been started, we studied the possible use of phenylglyoxal in the test for the intracellular bacterial killing capacity of PMN. Generally, phenylbutazone is used for this purpose (Solberg, 1972). Together with extracellularly working bactericidal agents, this inhibitor of the oxidative metabolism in PMN is added to the zero-time sample some minutes after the addition of live micro-organisms to the cells. In this way, the number of live bacteria in the PMN at the start of a killing experiment can be assessed. The number of live S. aureus found in phenylbutazone-treated PMN was about 30% higher than in phenylglyoxal treated cells. Neither phenylbutazone nor phenylglyoxal had any effect on the growth rate of S. aureus in the absence of PMN. Phenylbutazone inhibited the STZ stimulated oxygen consumption as well as did phenylglyoxal, but, as Fig. 5 shows, phenylbutazone did not inhibit the uptake of S. aureus after the start of phagocytosis as effectively as did phenylglyoxal (cf. Kj0sen, Bass0e & Solberg, 1976).

Phenylglyoxal is thought to act exclusively on the cell surface. Stein & Berestecky (1975) have shown that pronase treatment removes 74% of [¹⁴C]phenylglyoxal from HeLa S-3 cells. Moreover, only 7·7 % of [¹⁴C]phenylglyoxal was found in the 27000 g supernatant of the cells. Phenylglyoxal, however, may have bound to intracellular structures that sediment at that centrifugal force. Thus, in our opinion, this evidence is not absolute proof that phenylglyoxal reacts exclusively with cell-surface moieties. Yin (1975) found that lactate dehydrogenase was sensitive to phenylglyoxal: in a cell lysate treated with 150 µg phenylglyoxal/ml, only 47% of the original activity was found. Since no effect of phenylglyoxal on lactate dehydrogenase activity was observed in lysates made from phenylglyoxal-treated, washed cells she concluded that PMN, like HeLa cells (Stein & Berestecky, 1975), are probably impermeable to phenylglyoxal.

However, our finding that STZ-stimulated oxygen consumption by PMN was inhibited for about 90% by the addition of phenylglyoxal can hardly be explained solely on the basis of inactivation of cell surface components. Although it seems prob able that the oxygen-reducing complex of PMN, which generates superoxide, is located in the plasma membrane (Baehner, 1975; Goldstein, Cerqueira, Lind & Kaplan, 1977; Goldstein et al. 1975; Johnston et al. 1975; Roos et al. 1976; Roos, van Schaik, Weening & ‘‘‘ever, 1977b;Salin & McCord, 1974; Tanaka & O’Brien, 1975), it must be assumed that after 7 min of phagocytosis (see Results) a substantial part of this process takes place in the (membranes of the) phagosomes. Phenylglyoxal does not inhibit the oxygen consumption by affecting only the particle-cell association, because the PMA-stimulated oxygen consumption and lysozyme release were also strongly inhibited by phenylglyoxal.

Another indication for phenylglyoxal entry into the PMN was obtained by our studies on the glutathione levels. We found that phenylglyoxal substantially reduced the GSH level in intact PMN. This reaction is probably catalysed by glyoxalase 1 (Flohe & Gi.inzler, 1976; Schubert, 1935), because it depends on a cell constituent and divalent cations. These results are in agreement with those obtained by Amy & Rebhun (1977) with sea-urchin eggs. The inhibition of the PMA-stimulated oxygen consumption by phenylglyoxal was neither prevented nor released by EDTA; there fore, phenylglyoxal cannot act exclusively by glutathione binding.

Our experiments have excluded direct damage of the cells by phenylglyoxal: no lactate dehydrogenase release, no effect on ATP levels (which is contradictory to an effect on the energy metabolism) and no loss of viability was apparent. Moreover, many effects of phenylglyoxal on peritoneal macrophltges are reversible (Yin, 1975). In PMN, such reversibility could not always be detected, because PMN are too fragile for functional regeneration.

It is clear that cytochalasin B and phenylglyoxal exert their effects through different mechanisms. Although many of the observed effects of phenylglyoxal described in this paper resemble the previously described effects of cytochalasin B on endocytosis (Allison, Davies & de Petris, 1971; Axline & Reaven, 1974; Davies et al. 1971, 1973; Zigmond & Hirsch, 1972), cell mobility (Zigmond & Hirsch, 1972), membrane mobility (Allison et al. 1971; Axline & Reaven, 1974), transport of sugars (Estensen & Plagemann, 1972), and the stimulated oxygen metabolism (Roos et al. 1976), phenyl glyoxal and cytochalasin B differ in their effects on particle adherence (Allison et al. 1971), transport of nucleotides (Plagemann & Estensen, 1972), and exocytosis of lysosomal enzymes (Goldstein et al. 1975; Zurier, Hoffstein & Weissmann, 1973). The latter process, especially, shows a clear distinction; phenylglyoxal completely blocks lysosomal enzyme release stimulated by STZ (Figs. 3 and 4; Yin, 1975) or by chemotactic factors (Yin, 1975), whereas cytochalasin B actually facilitates this process (Davies et al. 1973; Goldstein et al. 1975; Yin, 1975; Zurier, Hoffstein & Weissmann, 1973).

Whereas cytochalasin B is thought to exert its action by disrupting microfilaments (Allison & Davies, 1974; Axline & Reaven, 1974; Carter, 1967; Spooner & Wessels, 1970), the mode of action of phenylglyoxal is still completely unknown. Inhibition of the association between the plasma membrane and cytoplasmic contractile elements (Oliver et al. 1976) seems likely. This would explain the effect of phenylglyoxal on endo- and exocytosis, on cell locomotion and on cap formation. Moreover, such interference might also inhibit particle binding, and - apart or as a consequence - generation of bactericidal oxygen metabolites. The observation that phenylglyoxal inhibited the respiration of PMN stimulated with STZ, as well as that of PMN incubated with IgG- or C3-coated zymosan, indicates that phenylglyoxal did not specifically act on either one of the membrane-binding places.

The inhibition of phagocytosis by phenylglyoxal cannot be explained by the effect of this agent on cell-particle adherence. In the presence of 100 /lg phenylglyoxal/ml, phagocytosis of STZ was inhibited by more than 90%, whereas rosette formation was decreased by only 50%. Moreover, Yin (1975) found that the uptake of polyvinyl toluene particles by mouse peritoneal macrophages was 70 % inhibited by 100 /lg phenylglyoxal/ml, while the particles remained adherent to the cell surface (light microscopic observation). Therefore, interference of phenylglyoxal with particle ad herence to the cell may play a role, but a direct action of phenylglyoxal on the phago cytic apparatus must be assumed to explain the complete inhibition of particle ingestion.

Thus, our studies have excluded the use of phenylglyoxal as a specific inhibitor of phagocytosis: at a dose of 40 µlg phenylglyoxal/ml, phagocytosis was inhibited by about 75 %, but chemotaxis, oxygen consumption and lysosomal enzyme release were also strongly inhibited. This property of phenylglyoxal can be put to use very well, however, in the bacterial killing test of PMN. In this test, a compound is needed to inhibit the uptake and intracellular killing of bacteria after a short initial ingestion phase. As such, phenylbutazone is used routinely (Solberg, 1972). We have found in the present study that phenylbutazone did not stop phagocytosis very effectively. Therefore, during the incubation with phenylbutazone, extracellular bacteria that have not yet been killed by the antibiotics can still be ingested and may increase the number of live, intracellular bacteria. Phenylglyoxal is much more effective than phenylbutazone in blocking the uptake of bacteria, even after the start of this process (Fig. 5). Indeed, a higher number of live, intracellular bacteria was found in phenyl but a zone-treated PMN than in phenylglyoxal-treated cells. For this reason, phenyl glyoxal is to be preferred in the bacterial killing test.

This study was in part financed by the Netherlands Organization for the Advancement of Pure Research (Z.W.O.), grant no. 91-5.

We thank Dr Janet Oliver for her valuable comments.