Eur J. Biochem 197, 127-133 (1991) c, FEBS 1991

O01429S691002170

3-Aminobenzamide inhibits cytotoxicity and adhesion of phorbol-ester-stimulated granulocytes to fibroblast monolayer cultures Thomas MEYER, Helgard LENGYEL, Werner FANICK and Helmuth HiLZ Institut fur Physiologische Chemie, Universitat Hamburg, Federal Republic of Germany

(Received September 24/November 22, 1990) - EJB 90 1138

Damage of 3T3 fibroblasts as induced by short-term co-cultivation with 0; -producing granulocytes, stimulated by 12-O-tetradecanoy1-phorbo1-13-acetate (TPA), was compared with that induced by treatment with enzymically generated 0, and with the alkylating agent dimethyl sulfate. The action of stimulated granulocytes was different in several aspects : (a) DNA fragmented by the products of TPA-stimulated granulocytes showed a biphasic alkaline elution pattern while fragmentation induced by alkylation or by enzymically produced 0; was monophasic. (b) Poly(ADP-ribosy1)ation of nuclear proteins after treatment with TPA-stimulated granulocytes exhibited a lag phase and was, in most experiments, less pronounced than after equitoxic dimethyl sulfate treatment. (c) 3- Aminobenzamide, the most widely used inhibitor of ADP-ribosylation, partially protected target cells from the cytotoxic effects of TPA-stimulated granulocytes, while it enhanced alkylation-induced and 0;-induced cytotoxicity. Protection by 3-aminobenzamide in the granulocyte system was apparently not mediated by an inhibition of nuclear poly(ADP-ribosy1)ation. Other inhibitors, like benzamide and nicotinamide, augmented cytotoxicity of TPA-stimulated granulocytes. The unique effect of 3-aminobenzamide in this system appeared to relate to TPA-induced adhesion of the neutrophils to surfaces. In the presence of 1 mM 3-aminobenzamide, but not of benzamide, the adhesion of stimulated granulocytes to 3T3 monolayer cultures was markedly reduced or even abolished. This effect was also seen in granulocyte preparations depleted of monocytes. Since 3- aminobenzamide at the doses applied does not inhibit TPA-induced superoxide production in isolated granulo- cytes, its specific anticytotoxic effect appears to result from a ‘dilution’ of granulocyte-derived damaging agents into the medium. Our data suggest that prevention of granulocyte adhesion is likely to reduce tissue damage and carcinogenesis in areas of chronic inflammation.

Growing evidence suggests that reactive oxygen species produced by phagocytic cells play a role in carcinogenesis [I, 21. They can cause cell injury not only to natural target cells, like microorganisms or tumor cells, but also to surrounding tissue cells, especially when exposed to a long-standing stimu- lation as seen in areas of chronic inflammation [3]. Superoxide anions and their products like H 2 0 2 or HOCl, which are released during the ‘oxidative burst’, can induce damage to membranes [3, 41. It has also been reported that 0; radicals are able to cause DNA fragmentation [5-91. The same has been shown for the radical-generating compound benzoyl per- oxide 17, 101, and for HzOz [ I l , 121.

Fragmentation of DNA, visible expression of the cell’s attempts to repair the damage, is usually accompanied by an ADP-ribosylation of nuclear proteins [13 - 191. In conse- quence, increased ADP-ribosyltranferase activity was also seen after treatment of cells with benzoyl peroxide [lo], enzymically generated 0; radicals [20,21] and with H 2 0 2 [12j.

Correspondence to H. Hilz, Institut fur Physiologische Chemie, Martinistrasse 52, W-2000 Hamburg 20, Federal Republic of Germany

Abbreviations. DME medium, Dulbecco’s minimal essential medium; TPA, 12-0-tetradecanoyl-phorbol-13-acetate.

Enzymes. Protein kinase C (EC 2.7.1.37); xanthine oxidase (EC 1.1.3.22); ADP-ribosyltransferase (EC 2.4.2.30); superoxide dismu- tase (EC 1.15.1.1).

Modification of nuclear proteins by poly(ADP-ribosy1)ation appears to be involved in the regulation of DNA repair, since inhibition of poly(ADP-ribosyl) polymerase activity po- tentiates cytotoxicity of alkylating agents [13, 14,22-241 and enhances their transforming capacity in several systems [24 - 271. On the other hand, it was also reported that cell death of a macrophage-like tumor cell line induced by H 2 0 2 treatment could be prevented by 3-aminobenzamide [ 121. However, op- posite data on the effect of 3-aminobenzamide in H,02-in- duced cytotoxicity have also been reported [28].

Because of these unclear actions of ADP-ribosylation in- hibitors when combined with agents known to arise from stimulated granulocytes, and in view of the possible impact of these reactive oxygen species on tumorigenesis in vivo (cf. [29 - 31]), we decided to study in co-cultivation experiments the effect of stimulated granulocytes on fibroblasts cultures as modulated by ADP-ribosylation inhibitors. We also compared the action of the phagocytic cells on DNA fragmentation and ADP-ribosylation with that of enzymically generated 0; radicals and of an alkylating agent. Here we demonstrate that the response of fibroblasts to the products of stimulated granulocytes is divergent from that evoked by cell alkylation and by 0; produced enzymically. We also show that 3- ammobenzamide, the compound most widely used to block poly(ADP-ribose) polymerase in vivo, differs from other in- hibitors by protecting rather than potentiating the cytotoxic

128

action of granulocytes stimulated by 12-0-tetradecanoyl- phorbol-13-acetate (TPA).

MATERIALS AND METHODS

Chemicals

TPA, 3-aminobenzamide, 3-aminobenzoic acid, hypo- xanthine and xanthine oxidase were obtained from Sigma (Deisenhofen); benzamide, benzoic acid, and NAD from E. Merck (Darmstadt), Dulbecco's mimimal essential (DME) medium and fetal calf serum from Boehringer (Mannheim).

Cell culture

Mouse 3T3 cells (NIH) were cultured in DME medium supplemented with 10% heat-inactivated fetal calf serum.

Human granulocytes were prepared according to [32] from fresh blood of healthy donors starting with the leukocyte- enriched fraction of the blood (buffy coat). The ability of each granulocyte preparation to produce superoxide anions in response to TPA treatment was tested using the cytochrome c assay [33].

Cytotoxicity test

3T3 cells were seeded at a density of 5 x lo4 cells/well in 12-well plates; 6 h after seeding the medium was removed and the cells were incubated with 1 ml DME medium containing 0.002% Tween 80, or 0.5% bovine serum albumin, instead of serum. Further additions are described in the individual figure legends. When indicated, granulocytes were added to final concentrations of 2- 10 x lo5 cells/ml. TPA was dissolved in acetone and added to a final concentration of 25 ng/ml. Sol- vent concentration was kept at 0.1%. After 30-60 min of treatment the wells were washed twice with fresh medium, and the cells were grown for 2 - 4 days in DME medium containing 10% fetal calf serum and the nicotinamide analogs when indi- cated.

Viability tests (colony forming ability) were performed in some of the experiments. Cells were trypsinized after treatment and seaded at 200 cells/well. Colonies (>20 cells) were counted after five days (triplicate determinations). The values closely related to the data obtained with the proliferative capacity test (cytotoxicity test) above.

Quantification of pory (ADP-ribose)

3T3 cells were grown in log phase for 3 - 4 days in 15-cm dishes. Before reaching confluence, medium was removed and the cells were treated in DME medium containing 0.002% Tween 80 instead of serum for the indicated times with TPA, TPA and granulocytes, of dimethyl sulfate in a total volume of 15 ml/dish. All incubations were carried out at 37°C in an air/CO, (95 : 5 ) atmosphere. Following incubation, the culture medium was removed and the dishes were treated with 3 ml ice-cold 20% trichloroacetic acid. The cells were scraped and collected, and the dishes were washed with another 3 ml20% trichloroacetic acid. The combined suspensions were homogenized in a glass/glass homogenizer, left on ice for 15 min and centrifuged for 5 min at 2500 g. The peIlets were washed twice with cold 20% trichloroacetic acid, twice with ethanol and subsequently dissolved in 0.1 ml100 mM NaOH. From this solution, DNA was determined according to [34],

and poly(ADP-ribose) as described previously using the phos- phoribosyl-AMP radioimmunoassay 135, 361.

Quantification of NAD

NAD was determined in the acid-soluble fraction (see above). The acid solution was extracted five times with a sevenfold volume of water-saturated ether. The remaining so- lution (pH 4- 5) was freed from ether by gentle heating and shaking. NAD was quantified using the cycling test described by Jacobson et al. [37].

Determination of DNA.fragmentation

3T3 cells were grown in 10-cm dishes under the conditions described above. Before treatment they were labeled 20 h with [I4C]thymidine (0.02 pCi/ml medium). Treatment of the cells was performed as described in the cytotoxicity test. DNA strand breaks were measured using the alkali elution pro- cedure of Kohn 1381.

Adhesion of granulocytes to 3T3 cultures

Adhesion was usually studied in DME medium containing 0.5% bovine serum albumin or 8% fetal calf serum. 3T3 cell monolayer cultures (0.5-2.0 x 105/well) were incubated for 30 - 60 min with 2 - 10 x lo5 granulocytes suspended in 1 ml medium. In some experiments, granulocyte suspensions were transferred to 15-cm dishes and kept for 60 min at 37°C (5%/ COz atmosphere) to allow monocytes to attach to the surface. The cells remaining in the supernatant were counted and used.

Treatment with enzyrnically generated 0;

DME medium of 3T3 cultures was replaced by medium containing 0.5% bovine serum albumin instead of serum. 0; was generated by the addition of hypoxanthine (25- 100 pM) and xanthine oxidase (50 mU/ml) which was added last,

RESULTS

DNA fragmentation induced by TPA-stimulated granulocytes versus fragmentation after treatment with dimethyl sulfate or superoxide anion

Freshly prepared human granulocytes react very sensi- tively to TPA. They are stimulated to produce 0; radicals by phorbol ester concentrations as low as 0.5 ng/ml (cf. [29]). First, we analyzed the cytotoxic and genotoxic potential of these superoxide anions and their products by exposing 3T3 cells to granulocytes for 60 min. Co-cultivation with granulo- cytes alone had no significant effect on the proliferative ca- pacity of the target cells (Table 1). The same was true when 3T3 cells were treated with TPA alone. However, when the granulocytes were stimulated to produce superoxide anions by the addition of TPA, a dramatic loss of proliferative ca- pacity of the 3T3 cells was encountered, thus confirming pre- vious observations based on the release of "Cr from pre- labeled ceIls [3, 4). The cytotoxic effect of TPA-treated gra- nulocytes could be prevented to a large extent by a combi- nation of superoxide dismutase and catalase, suggesting that most of the cytotoxicity was mediated by H 2 0 2 or by reaction products of H 2 0 2 (cf. [39, 91).

Clastogenic effects of TPA in the presence of leukocytes have been described by Birnboim [7] and by Cerutti et al. [8].

Table 1. Cytotoxic effect of TPA-.rtimulatedgranulocytes on 3T3Jibro- blasts and protection by superoxide dismutase and/or catalase NIH 3T3 cells (5 x 104/well) were treated for 60 min with granulocytes (Gran., 2 x 105/ml) and TPA (25 ng/ml) as indicated. Superoxide dismutase (SOD, 50 pg/ml) and catalase (50 pg/ml) were present during the 60-min incubation period where indicated. After washing the cultures with medium, cells were incubated in medium for three days, washed and counted. Cell growth was calculated as the increase in cell number after treatment. This parameter is an indicator of cell viability as shown by comparison with colony forming ability tests. Mean values + SD from three determinations. For further details see Methods

Additions Cell growth

A None TPA (25 ng/ml) Granulocytes Gram + TPA

B Gran. + TPA Gran. + TPA + SOD Gran. + TPA + catalase Gran. + TPA + SOD + catalase

% control

100 84k I

106 i 2 1 7 + 6

11 f 3 1 5 + 2 6 5 k 5 82 5 10

I I I I r I 2 L 6 8 1 0 1 2

f r ac t ion number Fig. 1. Alkaline elution profiles of 3T3 cell DNA after treatment of cells with dimethyl suIfLlte or with TPA-stimulated granulocytes. 3T3 monolyer cultures (5 x lo6 cells/lO-cm dish) were labeled for 12 h with ['4C]thymidine. After change of medium and further incubation for 6 h, medium was replaced by serum-free DME medium containing 10 pl acetone/lO mi medium(controls), or 25 ng TPA/ml, or 4 x lo6 granulocytes + TPA (25 ngiml), or 20 pM dimethyl sulfate (DMS) added immediately before incubation). After 30 min of incubation, cells were processed as described in Methods

These authors applied alkaline sucrose gradient centrifugation or alkaline unwinding to quantify DNA fragmentation. When we analyzed the DNA of 3T3 cells exposed to granulocytes in the presence of TPA, using the alkaline elution procedure 1381, an unusual elution pattern became evident (Fig. 1). While cells treated with the alkylating agent dimethyl sulfate showed the expected change in the slope of the elution curve, 3T3 cultures treated with granulocytes and TPA contained DNA that eluted in a biphasic manner. This type of elution suggested the presence of two forms of fragmented DNA, one more extensively degraded and the other consisting of only moder-

129

I I 150 4

U 30 60 90 100

incubation t i m e [ rn in ) Fig. 2. Accumulation of protein-bound poly (ADP-ribose) in 3T3 cells treated with dimethyl sulfate or TPA-stimulatedgranulocytes. 3T3 cells (5 x 106/10-cm dish; 10 ml medium) were treated with dimethyl sulfate (DMS, 200 pM), TPA (25 ngiml), granulocytes (1 x lo7), or TPA + granulocytes for the times indicated. Cells were then extracted on the dishes with cold trichloroacetic acid and analyzed as described in Methods

ately fragmented DNA. The biphasic elution pattern was seen after exposure to low and to high numbers of stimulated granulocytes. It is not clear at present whether this reflects two populations of damaged cells rather then two populations of differently damaged DNA in the same cells. It is important to note, however, that the highly fragmented DNA subpopu- lation does not seem to be derived from dead cells since vi- ability tests (dye exclusion) showed only a relative small (< loo/,) increase in damaged cells at the end of treatment (i.e. after a 60-min exposure). It is also important to note that enzymically generated 0; did not produce biphasic elution patterns. It rather lead to curves resembling those after dimethyl sulfate treatment (see later, Fig. 3). The clastogenic effect seen in the presence of TPA and granulocytes is me- diated by the combination of these two and not by one alone (cf. Fig. 1).

Poly ( A DP-ribosyl) ation of nuclear proteins in response to co-cultivution with stimulated granulocytes

Fragmentation of DNA in 3T3 cells exposed to stimulated granulocytes is accompanied by a marked augmentation of protein-bound poly(ADP-ribose) (Fig. 2). Nuclear poly- (ADP-ribosyl) residues started to rise after a lag phase of about 10 min. They reached a maximum near 60 min. The maximal value represents a 20-fold increase over controls, which is below (about two-thirds) the maximum obtained by treating the cells with about equitoxic concentrations of dimethyl sulfate (inhibition of proliferative capacity : 60 - 80%). Under both conditions, poly(ADP-ribosyl) groups slowly decreased during the second and third hour of incu- bation. Elevated poly(ADP-ribosy1)ation was accompanied by a partial depletion of the substrate NAD in both cases (not shown).

When ADP-ribosylated proteins were isolated from treated 3T3 cells by covalent chromatography on amino-

130

Table 2. Action of 3-arninobenzarnide on the proliferative capacity of 3T3 ceils treated with stiilaulated grunulocytes versus dirneth,yl sulfate (A) 3T3 cells (5 x 104/well; 1 ml DME containing 2% bovine serum albumin) were incubated with dimethyl sulfate (DMS 100 pM) and 3-aminobenzamide (2 mM) were indicated (hydrolysis of DMS at 37°C occurs within 30 min). Proliferative capacity was determined as the increase in cell number after two days. Average values + SD from triplicate cultures of a representative experiment. For details see Methods. (B) 3T3 cells (5 x 104/well) were exposed for 60 min to 2 x lo5 granulocytes and the additions as indicated (TPA = 25 ng/ ml; 3-aminobenzamide = 2 mM final concentration). Further treat- ment of cultures and determination of proliferative capacity (increase in cell number after two days) was performed as described in Methods. Average values from 4 - 6 independent experiments

Additions Relative proliferating capacity

A None 3-Aminobenzamide Benzamide DMS DMS + 3-aminobenzamide DMS + benzamide

B None TPA TPA + 3-aminobenzamide TPA + benzamide

100 88 k 3 85 f 6 30 f 1 13 f 2" 11 * 3"

100 40 +_ 7 13 + 6 b 27 f 5'

a Significance of difference to DMS: P I 0.001. Significance of difference to TPA: P I 0.002. Significance of difference to TPA: P 5 0.05.

phenyl boronate columns [18], the general pattern after treat- ment with granulocytes + TPA was comparable to that seen after alkylation, although total amounts were somewhat smaller than after dimethyl sulfate (cf. Fig. 2). Stimulated granulocytes tended to induce a preponderance mono(ADP- ribosy1)ation over poly(ADP-ribosy1)ation of acceptors (not shown).

3-Aminobenzumide as a protecting agent against granulocyte-derived cytotoxicity

Treatment of proliferating cells with monofunctional non-crosslinking alkylating agents produces dose-dependent cytotoxic effects which can be greatly potentiated by the ad- dition of inhibitors of poly(ADP-ribose) polymerase [14,22 - 241. Co-cytotoxicity of 3-aminobenzamide with dimethyl sulfate was also evident in 3T3 cells (Table 2A). However, when we analyzed the effect of 3-aminobenzamide on the action of TPA-stimulated granulocytes, no potentiation of cytotoxicity was seen. Instead, protection from the detrimen- tal effects of the granulocyte system was observed (Table 2B). 3-Aminobenzoate, which cannot block ADP-ribosylation, provided only marginal protection (not shown).

Although these data seemed to be consistent with a mech- anism of 3-aminobenzamide action that involved poly(ADP- ribosyl)ation, it was surprising that 3-aminobenzamide added after the treatment with granulocytes had practically lost its protective action, while it still enhanced dimethyl-sulfate-in- duced cytotoxicity.

Direct evidence against a link between protection against cytotoxicity and inhibition of poly(ADP-ribose) polymerase

came from comparative studies with additional polymerase inhibitors (Table 2 B). Benzamide (like nicotinamide) did not protect 3T3 cells from the cytostatic action of stimulated gra- nulocytes, but significantly enhanced inhibition of cell pro- liferation. The unique protecting action of 3-aminobenzamide was also evident when higher numbers of granulocytes were used, or when viability and proliferative capacity were tested by the colony forming ability of treated cells (see below).

The protecting effect of 3-aminobenzamide in the granulo- cyte system was also seen at the level of DNA fragmentation (Fig. 3), although in a number of experiments protection by 3-aminobenzamide was only marginal. Benzamide, on the other hand, always enhanced over-all fragmentation of DNA as induced by TPA-stimulated granulocytes. By contrast, no protective action of 3-aminobenzamide was observed in com- bination experiments with an alkylating agent (Fig. 3 B) or, more importantly, with enzymically generated 0; (Fig. 3 C). With these two clastogenic agents, 3-aminobenzamide rather enhanced fragmentation of DNA, as did benzamide and nic- otinamide.

Although nicotinamide and its analogs are able to inhibit 0; production in TPA-stimulated granulocytes (unpublished experiments), the concentrations required for 50% inhibition were 10-20 times higher than those effective in the co-cytotoxicity tests discussed above; there was no basic dif- ference between nicotinamide, benzamide and 3-amino- benzamide. Also, the observation that 3-aminobenzamide en- hanced rather than inhibited the clastogenic action of enzymically generated 0; radicals strongly indicated that the specific anticytotoxic effect of 3-aminobenzamide in the gra- nulocyte system was not the result of an inhibitory action on 0; production. We than compared the effect of 3-aminobenzamide on the cytotoxicity of enzymically gener- ated superoxide anion radicals with that of granulocyte-de- rived 0; (Table 3). Damage imposed by treatment with hypoxanthine/xanthine oxidase could not be reduced by 3- aminobenzamide. Rather, the amino derivative acted in this system like benzamide, enhancing cytotoxicity of 0;. The divergent effect on cytotoxicity of 3-aminobenzamide in the two systems was evident from the analyses of proliferation rates (not shown) as well as of plating efficiences (Table 3).

3-Aminobenzamide as a specific inhibitor of TPA-induced adhesion of granulocytes to 3T3 cell cultures

The findings presented in the preceding section indicated that the protective action of 3-aminobenzamide was restricted to the granulocyte system and that it occurred somewhere between the TPA-induced production of cytotoxic agents and their reaction with the target cells.

Recent observations suggest that stimulation of neutro- phils with chemotactic agents or with calcium ionophores induces their adherence to surfaces (cf. [40]). Attachment to endothelial cells could also be provoked by activators of neutrophil protein kinase C [40 -431. When we incubated 3T3 cultures with granulocytes in the presence of TPA for 15- 60 min, a large fraction of the granulocytes became firmly attached to the cultures, their number being of the same order of magnitude as the number of the 3T3 monolayer cells. At- tachment was not significantly changed when the granulocytes were depleted of monocytes, or when the medium contained 0.5% bovine serum albumin instead of serum. Incubation in the presence of 1 - 2 mM 3-aminobenzamide strongly reduced the number of adhering granulocytes (60 or 120 min of incu-

131

2 L 6 8 1 0 Fraction number

1 'SZ

'0 5 1 DMS+ABam

2 1 6 8 1 0 Fraction number

X D 4 ABarn I = l o 0

2 1 6 8 1 0 Fraction number

Fig. 3. Influence of3-aminobenzumide on DNA fragmentation induced in 3T3 cells with different treatments. 20 x lo6 3T3 cells were labeled with ['4C]thymidine and processed in DME containing 2% bovine serum albumin, as described in the legend to Fig. 1. (A) Treatment with 1 x lo6 granulocytes/dish for 60 min with 25 ng/ml TPA and 2 mM 3-aminobenzamide (ABam), or 2 mM benzamide (Barn) where indicated. (B) Treatment for 60 min with 25 pM dimethyl sulfate (DMS) added in 10 p1 acetone. (C) Treatment for 60 mln with 100 pM hypoxanthine, 50 mU/ml xanthine oxidase (xo) and 3-aminobenzamide (ABam) where indicated

Table 3. Plating ej'ciency of 3T3 cells after treatment with enzyme- derived versus granulocyte-derived 0; as modified by 3-amino- henzamide 3-Aminobenzamide (ABam) was added to 2 mM final concentration, xanthine oxidase/hypoxanthine (XO/HX) as described in Methods. Mean values f S D from 3 - 5 experiments

Additions Relative plating efficiency ~~ -

None ABam XO/HX XO/HX + ABam Gran. + TPA Gran. + TPA + ABam

100 f 22 9 9 + 6

4 6 + I 57 PIO.01

34 ' P I 0 . 0 0 5 4 9 * 5

bation; cf. Fig. 4). Microscopic counting indicated that the fibroblasts remained attached under these conditions.

The action of 3-aminobenzamide appeared to be specific since no significant influence was seen in the presence of benzamide, while 3-aminobenzoic acid exerted only marginal protection (not shown). Thus, the specificity of action of the nicotinamide analogs on TPA-induced granulocyte adhesion follows the same pattern as their effect on granulocyte-induced cytotoxicity.

DISCUSSION

The cytotoxic action of alkylating agents, as well as that of the superoxide/H202 system, is thought to be mediated at least in part by their genotoxic effects [44, 451. Both induce fragmentation of DNA and both provoke the modification of nuclear proteins by ADP-ribosylation (cf. [12, 14, 44-46]). The oxidative burst as elicited in granulocytes by chemotactic agents or TPA also leads to the formation of 0,. It was anticipated, therefore, that the action of stimulated granulo- cytes on neighboring cells induces the same pattern of response as seen after treatment with enzymically generated 0; or with monofunctional non-crosslinking alkylating agents.

A comparative analysis, however, revealed that the dam- age induced by stimulated granulocytes differed significantly

TPA : - + - - + + ABam : - - + - + - Barn : + - + - - -

Fig. 4. Influence of 3-aminobenzamide versus benzarnide of TPA-in- duced adhesion ofgranulocytes to 3T3 monolayer cultures. 2 x lo5 NIH 3T3 cells were seeded in 12-well plates and incubated (37°C 5% COz) in 2 ml DME mediumcontaining 8% heat-inactivated fetal calfserum. After 6 h, medium was removed, and 6 x lo5 human granulocytes, suspended in 1 ml DME containing 2% bovine serum albumin, were added. Granulocytes had been depleted of monocytes by prior incu- bation (1 h, 37"C, 5% COz) of the cell suspension in plastic dishes. All wells received 2 p1 acetone or TPA in acetone as indicated. The amides were added as concentrated aqueous solutions. After 60 min at 37"C, the granulocyte suspension was removed, and the wells were carefully rinsed three times with agitation, using 2 ml phosphate- buffered saline. The remaining cells were suspended in Isoton 11 (Coulter Electronics) and counted microscopically or in a Coulter counter. Adhering cells are defined as the difference between controls (no granulocytes added) and the cells remaining after rigorous washings

from that provoked by the chemicals. The differences pertained to the clastogenic effects as characterized by a bi- phasic elution pattern of DNA, and to the kinetics and pattern of ADP-ribosylation. The most important difference, how- ever, was seen when inhibitors of ADP-ribosylation were ap- plied in conjunction with the cytotoxic agents: while benza- mide and nicotinamide acted as co-cytotoxics with all agents applied, 3-aminobenzamide did not. It rather revealed a basic difference between granulocyte-derived noxious agents and enzymically generated superoxide. In the granuiocyte system

132

this ADP-ribosylation inhibitor acted as a potent protecting agent, preventing to a large extent the TPA-induced cytotoxic effects. In most experiments, it also reduced the fragmentation of DNA in the co-cultivated cells.

This specific action of 3-aminobenzamide was apparently not mediated by an inhibition of ADP-ribosylation because other, more potent, inhibitors did not provide protection but enhanced cytotoxicity. This was also true for 3-amino- benzamide when combined with enzymically generated 0; or with alkylating agents. Furthermore, 3-aminobenzoic acid which is not an inhibitor of ADP-ribosylation, also exerted a slight protective effect in the granulocyte system (unpublished results).

3-Aminobenzamide at the effective concentrations did not inhibit 0; production of granulocytes. ICs0 values for 0; production were 10 - 20 times higher than those providing growth protection (unpublished experiments). The effect, therefore, had to pertain either to a type of cytotoxicity not related to the effects of reactive oxygen species (e. g. proteases), or protection by the amide concerned the mode of delivery of damaging agents from the granulocytes to the target cells. That reactive oxygen rather than proteases confer cytotoxicity to the co-cultivated 3T3 cells was shown by the prevention of most of the damage by the addition of superoxide dismutase/ catalase (Table 1). Protection, therefore, was likely to relate to the mode of delivery of 0; by the granulocytes.

Due to the presence of detoxifying factors in biological fluids or in culture media, effective delivery of the noxious agents presumably requires close proximity or attachment of stimulated granulocytes to target cells. Attachment has indeed been shown to occur in the case of endothelial cells [40-431 and other cell types [47]. In our system, we could demonstrate that granulocytes under the influence of TPA firmly attach to 3T3 monolayer cultures. In confluent 3T3 cultures, a signifi- cantly lower number of granulocytes adhered in response to TPA. The attachment could be specifically prevented by 3- aminobenzamide, not by benzamide, and only slightly by 3- aminobenzoate. This is the same specificity as seen in the protection against granulocyte cytotoxicity. It therefore pro- vides a reasonable explanation for that effect: direct cell - cell contact between granulocytes and target cells may guarantee delivery of ‘undiluted’ damaging agents, while prevention of the attachment by 3-aminobenzamide should result in a di- lution of the noxious agents into the entire medium leading to an apparent protection’.

In this context it might be of interest to note that ‘natural’ stimulation by a chemotactic peptide (Met-Leu-Pro) of gra- nulocytes co-cultivated with 3T3 cells did not lead to cytotoxicity (unpublished experiments). Presumably, early down-regulation of the receptor-mediated effect prevents the accumulation of concentrations of reactive oxygen species high enough to overcome the scavenger systems in the medium and in the target cells.

The mechanism of action of 3-aminobenzamide on gra- nulocyte adhesion is not known. However, the specific effect of the amide may help to elucidate the reactions leading from TPA-induced activation of granulocytic protein kinase C to the docking of phagocytic cells. Our data also suggest that adhesion of defense cells as induced by the tumor promoter TPA may be an important factor in endogenous carcino- genesis.

This work was supported by a grant from the Deutsche Forsc/2ungsgemeinsc~iaft SFB 232 We thank K. Klapproth and K. Miiller for carrying out some determinations and for providing cell

cultures, and B. Nippa and R. Pforte for typing the manuscript. The provision by Prof. Kiihnl (Bluttransfusionsmedizin. Universitats- Krankenhnus Eppendorf) of a fraction of enriched human leukocytes is gratefully aknowledged.

REFERENCES 1.

2. 3.

4.

5.

6. 7. 8. 9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

Demopoulo, H. B., Peitonigro, D. D. & Seligman, M. L. (1983)

Roman-Franco, A. A. (1982) J . Theor. Biol. 97, 543-555. Nathan, C. F., Silverstein, S. C., Bruckner, L. A. & Cohn, Z. A.

Weiss, S. J., Young, J., LoBuglio, A. F., Slivka, A. & Nimeh, N.

Brawn, K. & Fridovich, I. (1981) Arch. Biochem. Biophys. 206,

Weitzmann, S . A. & Stossel, T. P. (1981) Science 212, 546-548. Birnboim, H. C. (1982) Science 215, 1247-1249. Emerit, I. & Cerutti, P. A. (1983) Nature 293, 144-146. Shacter, E., Beecham, E. J., Covey, J. M., Kohn, K. W. & Potter,

M. (1 988) Curcinogenesis 9, 2297 - 2304. Hilz, H., Koch, R., Kreimeyer, A., Adamietz, P. &Jacobson, M.

K. (1985) in ADP-ribosylation of proteins (Althaus, F., Hilz, H., Shall, S., eds) pp. 518-525, Springer-Verlag, Berlin, Heidelberg, New York.

Hoffmdnii, M. E. & Meneghini, R. (1979) Photoche~. Photohioi. 30, 151 -155.

Schraufstatter, I. V., Hyslop, P. A,, Hinshaw, D. B., Spragg, R. G., Sklar, L. A. & Cochrane, C. W. (1986) Proc. Natl Acad. Sci. USA 83,4908-4912.

Berger, N. A., Sikorski, G. W., Petzold, S. J. & Kurohara, K. K. (1979) J . Clin. Invest. 63, 1164-1171.

Durkacz, B. W., Omidiji, O., Gray, D. A. & Shall, S. (1980) Nature 283, 593 - 596.

Benjamin, R. C. & Gill, D. M. (1980) J . Biol. Chem. 255,10502- 10508.

Wielckens, K., Schmidt, A., George, E., Bredehorst, R. & Hilz, H. (1982) J . Bid. Chem. 257, 12872-12877.

Jacobson, E. L., Antol, K. M., Juarez-Salinas, H. & Jacobson,

Kreimeyer, A., Wielckens, K., Adamietz, P. & Hilz, H. (1984) J .

Adamietz, P. & Rudolph, A. (1984) J. Biol. Chem. 259, 6844-

Carson, D. A., Seto, S. & Wasson, D. B. (1986) J . Exp. Med. 163,

Singh, N., Poirier, G. & Cerutti, P. (1985) Biochern. Biophys. Res.

Nduka, N., Skidmore, C. J. & Shall, S. (1980) Eur. J . Biochern.

Hilz, H. (1981) Hoppe-Seyler’s Z. Physiol. Chem. 362, 1415- 1425.

Lubet, R. A., McCarvill, J. T. & Schwarz, J. L. (1985) in ADP- rihosylation qfprotein (Althaus, F., Hilz, H., Shall, S., eds) pp. 471 - 474, Springer-Verlag, Berlin, Heidelberg, New York.

Okamoto, H. & Yamamoto, H. (1983) in ADP-rihosylution, DNA repair and cancer (Miwa, M., Hayaishi, O., Shall, S., Smulson, M. & Sugimura, T., eds) pp. 297-308, Japanese Scientific Society Press, Tokyo.

Miwa, M., Ishikawa, T., Kondo, T., Takayama, S. & Sugimura, T. (1985) in ADP-ribosylution ofproteins (Althaus, F., Hilz, H., Shall, S., eds) pp 480 -483, Springer-Verlag, Berlin, Heidelberg, New York.

Kasid, U. N., Stefanik, D. F., Lubet, R. A,, Dritschilo, A. & Smulson, M. E. (1986) Carcinogenesis 7, 327-330.

Cantoni, O., Murray, D. & Meyn, R. E. (1986) Biochim. Biophys. Acta 867, 135-143.

Goldstein, B. D., Witz, G., Amoruso, M., Stone, D. S. & Troll, W. (1981) Cancer Lett. 11, 257-262.

Zimmermann, R. & Cerutti, P. (1984) Proc. Natl Acad. Sci. USA

J . Am. Cell. Toxicol. 2, 173- 184.

(1979) J . Exp. Med. 149, 100-113.

F. (1981) 1. Clin. Invest. 68, 714-721.

414-419.

M. K. (1983) J . Biol. Chem. 258, 103-107.

Biol. Chem. 259,890 - 896.

6846.

746 - 751.

Commun. 126, 1208-1214.

105, 525 - 530.

81,2085-2087.

133

31.

32.

33.

34. 35.

36.

37.

38.

39.

Lewis, J. G. & Adams, D. 0. (1986) Cancer Res. 46, 5696-

Claesen, H. E. & Feinmark, S. J. (1984) Biochim. Biophys. Actu

Cohen, H. J.&Choraniec, M. E. (1978)J. Clin. Invest. 61,1081-

Burton, K. (1968) Method.7 Enzymol. 12, 163-166. Wielckens, K., George, E., Pless, T. & Hilz, H. (1983) J . Bid.

Wielckens, K., Bredehorst, R. & Hilz, H. (1984) Methods

Jacobson, E. L., Lange, R . A. &Jacobson, M . K. (1979) J . Cell.

Kohn, K. W., Erickson, L. C., Ewig, R. A. G. & Friedman, C.

Simon, R. H., Scoggin, C. H. & Patterson, D. (1981) J . Bid.

5700.

804, 52 - 57.

1087.

Chem. 258,4098-4104.

Enzymol. 106,472-482.

Physiol. 99, 417 -426.

A. (1976) Biochemistry 15, 4629-4631.

Chem. 256,7181 -7186.

40.

41.

42.

43.

44.

45. 46.

47.

Meurer, R. & Mac Intyre, D. E. (1989) Biochem. Bioph-ys. Re.s.

Lo, S . K., Detrners, P. A., Levin, S. M. & Wright, S. D. (1989) J .

Ley, K., Lundgren, E., Berger, E. & Arfors, K. E. (1989) Blood

Skoglund, G., Patarroyo, M., Forsbek, K., Nielson, K. & In-

Boesen, E. & Davis, W. (1969) Cytotoxic drugs in the treatment

Meneghini, R. (1988) Mutat. Res. 195, 215-230. Althaus, F., Hilz, H. & Shall, S. (1985) ADP-ribosylation ofpro-

Brown, P. J. (1988) Biochem. Biophys. Res. Commun. 155,603-

Cornmun. I59, 763 - 769.

EX^. Med. 169,1719-1793.

73, 1324 - 1330.

gelman-Sundberg, M. (1988) Cuncer Res. 48,3168-3172.

Of cancer, E. Arnold, London.

teins, Springer-Verlag, Berlin, Heidelberg, New York.

607.