Acetylsalicylic acid (ASA) protects the prostaglandin-cAMP-system of human hypernephroma cells against irradiation-induced alterations.

There is abundant evidence that inhibitors of prostaglandin (PG) biosynthesis might increase the radioresponse of certain tumour cells. This study investigated specific PG binding sites, eicosanoid production as well as intracellular cAMP levels in cultured human hypernephroma cells derived from 11 patients upon nephrectomy. Scatchard analyses of the binding data revealed specific PGE1-, PGE2- as well as PGI2-binding sites (PGE1: Bmax = 755 +/- 206 fmol mg-1 protein, Kd = 3.7 +/- 2.7 nM PGE2: Bmax = 494 +/- 221 fmol mg-1 protein, Kd = 4.2 +/- 2.5 nM; PGI2: Bmax = 693 +/- 164 fmol mg-1 protein, Kd = 6.0 +/- 4.5 nM). Significant (P < 0.01) increase in PG binding sites expressed on human hypernephroma cells (PGE1: Bmax = 1084 +/- 303 fmol mg-1 protein, Kd = 2.8 +/- 1.3 nM; PGE2: Bmax = 663 +/- 309 fmol mg-1 protein, Kd = 2.2 +/- 1.5 nM; PGI2: Bmax = 1021 +/- 391 fmol/protein, Kd = 4.2 +/- 3.6 nM) and inhibition of PG biosynthesis (TXB2: -82.5%, PGE2: -87.5%. PGD2: -80.6%, PGF2: -81.3%) were found after acetylsalicylic acid (ASA)-treatment (0.5 mg 10(-6) cells for 24 h). Following irradiation (60Co, 1.0 Gy/min-1 over 10(min), PG binding sites (PGE1: Bmax = 266 +/- 153 fmol mg-1 protein, Kd = 5.0 +/- 5.0 nM; PGE2: Bmax = 148 +/- 66 fmol mg-1 protein, Kd = 4.7 +/- 3.6 nM; PGI2: Bmax = 325 +/- 194 fmol mg-1 protein, Kd = 6.8 +/- 7.1 nM) were significantly (P < 0.01) diminished. However, irradiation had no significant effect on PG binding sites in ASA-pretreated cells (PGE1: Bmax = 699 +/- 240 fmol mg-1 protein, Kd = 3.5 +/- 1.8 nM; iloprost: Bmax = 766 +/- 452 fmol mg-1 protein, Kd = 3.2 +/- 2.2 nM). Although there was no significant difference in the basal values for cAMP between control and ASA-treated group cells, the PG-induced cAMP-production was less pronounced in the control group. Taken together, the findings suggest that ASA may modify the radioresponse of cultured human hypernephroma cells by preventing the decrease of PG binding sites induced by irradiation.

Prostaglandins (PGs) are considered to play an important role in the regulation of tumour cell growth and metastases formation (Honn et al., 1981). In recent years several studies have shown that inhibitors of PG biosynthesis, including indomethacin, may improve the therapeutic effect of chemo- (Powles et al., 1978), immuno- (Chun & Hoffman, 1987) and/or radiotherapy (Furuta et al., 1988a) regimens of some tumours. However, whereas indomethacin remarkably increased tumour cell radioresponse, it had only minimal effect on the radioresponse of normal tissues such as hair follicles, jejunum or hematopoietic tissue (Furuta et al., 1988a). Furthermore, the response to indomethacin treatment was dependent on the ability of the tumour to produce PGs, mainly PGE2 and PGI2 (Furuta et al., 1988b). Apart from the fact that the mechanisms by which indomethacin potentiates tumour radioresponse are still unclear to date, it appears that an increased tumour response might be achieved by lowering PGs in the tumour. This, however, would imply that PGs are not radioprotective agents as has been suggested by Hanson & Ainsworth (1985) or Walden et al. (1987). Milas and coworkers (1990) have also found that potentiation of the tumour radioresponse induced by indomethacin is more significant when it is given after rather than before irradiation. Beside recent reports on the dependence of indomethacin-augmented radioresponse on immunocompetence of the tumour host (Milas et al., 1990), another possibility for PG action on tumour cells would be radiosensitisation by PGs. Moreover, one has also to bear in mind that the PG production may be remarkably different among various tumours (Malachi et al., 1981;Ziboh et al., 1981).
PGs exert their effects after interaction with specific cell surface receptors (Virgolini et al., 1992). We have demonstrated in human thyroid cancer (Virgolini et al., 1988) and hepatomas (Virgolini et al., 1989) that the number of PG receptors significantly (P<0.001) correlates with the cellular differentiation of the tumour. In these studies we found that high differentiated cancers seems to possess a higher PG receptor density than do anaplastic or less differentiated cancers. Furthermore, patients with a higher PG receptor density might have a better clinical prognosis. We have now used a similar receptor assay and have characterised the PG receptor on cultured human hypernephroma cells. These cells were shown to produce an increased amount of PGs of the E-series as compared with normal tissue (Cummings & Robertson, 1977). Furthermore, hypernephroma cells are relatively radiosensitive to high-dose radiation therapy (Halperin & Harisindis, 1983;Lang & deKernion, 1981). In a further step we wondered to what extent the PG system (PG receptors, cAMP-formation, PG production) would be influenced by irradiation, and investigated the effects of acetylsalicylic acid (ASA) on this system.

Materials and methods
Hypernephroma cell culture preparation Tissue specimens (approximately 1 cm3) were obtained intraoperatively from 11 patients aged 61 ± 12 years (nine males and two females) undergoing nephrectomy for renal cell carcinomas. Five of the patients had metastatic cancer, four were free of metastases. All patients had given written and informed consent to the study. From each patient a cell line was cultured. Renal cell carcinoma tissue specimen were made into a cell suspension by an enzymatic procedure as follows: suspended cells were incubated on a Petri dish in a humidified atmosphere of 2.5% CO2, 97.5% 02 at 37°C. The growth medium was changed three times a week. When the cells were confluent, a cell suspension was prepared by incubation with trypsin-EDTA-solution (0.05% trypsin, 0.02% EDTA) in phosphate buffered saline (PBS). Cells were routinely maintained in medium consisting of nutrient mixture F12 (Ham) supplemented with 12.5% fetal bovine serum, 0.24 mg collagenase, 0.01 mg DNA-se, 100 I.U. penicillin and 00 gg streptomycin ml-'. The cells were plated into 6 cm plastic Petri dishes and stored in the incubator for 4-18 days for attachment and growth. When the Petri dishes were densely covered with cells, a single-cell suspension was prepared. One part of the cell suspension was plated for further propagation, another part was used for testing the multiplication of cells with different concentrations of Cytochalasin B (CB) and one part was plated into plastic tissue-culture chambers (Lab-Tek), fixed after 24-72 h and checked for content of cytokeratin as an indicator of malignancy (Diereck et al., 1991). If there were only cytokeratin-positive cells, the cell culture was used for combined testing in the second passage. Cultures which were contaminated with cytokeratin-negative cells were further propagated and checked again when the cells became confluent. If there were still cytokeratin-negative cells after the fourth passage, the cell line was excluded from the test.
Cells from each cell line (one from each patient) were divided into four groups: (1) Control group; (2) ASA-treated group: cells were cultured in the presence of ASA (Bayer, Leverkusen, Germany; concentration 0.5 mg-6 cells ml-' medium for 24h); (3) Control group irradiated: irradiation was performed with gamma rays from a 'Co unit (Gammatron, Siemens) with a source-surface distance of 75 cm and a dose rate of 1.0 Gy min-' over 10 min. Dose measurements (calibrations) were carried out with a standard dosimeter (Farmer 0.6 cm, Nucl. Enterprise). (4) ASA-treated group irradiated: the same irradiation scheme as in group 3. Cellular viability was assessed by phase-contrast microscopy and Trypan Blue exclusion criteria (Wandl et al., 1989).
Binding studies PG receptor binding studies were carried out according to the methodology described previously (Virgolini et al., 1992). All assays reported here were performed approximately 2 h after irradiation. Briefly, the cells (4-6 x 106 cells in each group) were washed twice in assay buffer containing 50 mM Tris-HCI (pH 7.5), 5 mM MgCl2, 1 mM CaC12, 0.1 M NaCI and centrifuged at 500g for O min at 4°C (Beckman J-6B Centrifuge, Munchen, Germany). The pellet was suspended in 4°C assay buffer at a protein concentration of approximately 300 yg membrane protein ml-' (200-420 fg ml-') as determined by the assay kit provided by Bio-Rad Laboratories (Coomassie Brilliant Blue G-250, Richmond, CA). In preliminary experiments the time course of binding as well as the dependency of binding on temperature was studied. Based on the results of these experiments all further incubations were performed at 4°C for 50 min to ensure equilibrium.
After incubation for 50 min at 4°C the reaction was diluted rapidly with 4 ml of 4°C assay buffer and filtered through a Whatman GF/B filter (Whatman Inc., Clifton, NY) under reduced pressure (-60 kPa). The filters were then dried, transferred into scintillation vials (Packard, Downers Grove, IL) and taken up into 10 ml scintillation fluid (Pico-Flour TM30, Packard). The radioactivity in the samples was counted for 5 min in a liquid scintillation counter (LKB Wallace, 1215 Rackbeta, Finland).

Measurement of cAMP formation
Four to 6 x 106 tumour cells in each group were washed twice in 50 mM Tris-HCI buffer (pH 7.5) and resuspended in 50 mM Tris-HCI buffer (pH 7.5, 4°C) containing theophylline (BYK, Gulden Konstanz, Germany, 120 mg 1') to block the phosphodiesterase and ASA (50 mg 1`) to prevent endogenous PG synthesis. Cells were incubated in a 37°C shaking water bath for 30 min with either PGE,, PGE2 or iloprost in a concentration range from I0-M to 10-9M. After 30min the cells were homogenised by ultrasound (Sonicator, W-220F, Plainview, NY) and ultraturrax (TP18/10, Janke & Kunkel GmbH, Staufen, Germany) for 10s. The incubation was stopped by centrifugation at 5000g for 10min at 4°C. The cAMP concentration in the supernatant was determined by RIA according to the manufacturer's description (Amersham). Briefly, the samples were mixed with 251I-cAMP and rabbit anti-cAMP serum and incubated for 3 h at 4°C. The antibodybound cAMP was then extracted with a donkey anti-rabbit antibody, which was coated onto magnetisable polymer particles. These were mixed and left to react with the antibodies for 10 min at 25°C. The antibody-bound fraction was obtained with a magnetic separator (Amerlex-M Separator, Amersham) and the activity was determined by a gamma counter (Riastar, Packard). The intra-assay variability was 3.5 ± 1.3% and the inter-assay 5.6 ± 2.3%.
Saturation of 3H-PG binding to human hypernephroma cells was studied by incubating increasing concentrations of ligand in the absence and presence of an excess of the same unlabelled agonist. Binding was saturable and indicated a single class of high affinity binding sites for all three ligands within the ligand range studied (Figure 3a-b). After preincubation of the cells with ASA (group 2) a significantly (P<0.01) increased capacity to bind the PGs was found. The respective binding data are given in Tables I-III showing that ASA treatment increased 3H-iloprost receptors from 693 ± 164 to 1021 ± 391 fmol mg-' protein (P<0.05), 3H-PGEI-receptors from 755 ± 206 to 1084 ± 303 fmol mg-' protein (P<0.01) and 3H-PGE2-receptors from 494 ± 221 to 663 ± 309 fmol mg-' protein (P <0.05). Increase in the binding capacity was accompanied by a significant (P <0.05) decrease in the dissociation constant Kd.
Effect of PGs on cAMP-formation There was no significant difference in basal values between control (30.1 ± 10.5 pmol mg -') and ASA-treated group (28.0 ± 11.5 pmol mg-' protein). However, the basal values of control group irradiated (17.6 ± 8.9 pmol mg-' protein) and of ASA-treated group irradiated (15.8 ± 6.1 pmol mg-' protein) were significantly (P<0.001) lower as compared to the control group.
Iloprost, PGE, and PGE2 significantly (P<0.01) stimulated cAMP-production in all four groups dose-dependently. However, the half maximal effective doses (ED50) 70-1 I were significantly different between the four groups (Figure Conversion of exogenously added '4C-AA to its metabolites 4a-c). The corresponding ED_% values for iloprost were 12 ± 8 x Io-' M for the control group, 35 ± i 3 x l0-7 M for Hypernephroma cells converted exogeneous precursor AA to the ASA-treated group, 80 ± 21 x 10-7 M for the control a various number of compounds (HETE, PGD2, thromboxgroup irradiated, and 43 ± 11 x 10-7 M for the ASA-treated ane B2 (TXB2), PGE2, PGF2J) (Table IV). The main group irradiated. The ED50 for PGE, were 25 ± 8 x 10-5 M metabolites of AA conversion of cultured hypernephroma for the control group, 30 ± 10 x 10-7 M for the ASA-treated cells were PGE2 and TXB2 in both the control and the group, 11 ± 6 x 10-7 M for the control group irradiated, and irradiated groups. After ASA treatment, PG production was 88 ± 15 x 10-7 M for the ASA-treated group irradiated. The significantly decreasedthe main metabolite of AA convercorresponding ED50 values for PGE2 were 11 ± 4 x 10-5 M sion was HETE.
for the control group, 61 ± 22 x 10-8 M for the ASA-treated group, 86 32 x 10-7 M for the control group irradiated, Discussion and 33 ± 12 x 10-8 M for the ASA-treated group irradiated.
Previously, we have identified and described PG binding sites a on tumour cells (Virgolini et al., 1988(Virgolini et al., , 1989 and have demonstrated that cellular differentiation could be closely associated with receptor density and binding affinity. In this .' 600-study comparable low numbers of specific binding sites for 3H-PGEI, 3H-PGE2 as well as 3H-iloprost were found on cultured human hypernephroma cells which resemble the i, 400 , z _results obtained for thyroid cancer (Virgolini et al., 1988) and E human hepatocellular cancer (Virgolini et al., 1989). We -a Al / (Virgolini et al., 1988(Virgolini et al., , 1989 and others (Robertson et al.,!tn 200 ing sites in malignant tissues which might be due to an C g increased PG-production by malignant cells (Bennett et al., ffi 1982;Porteder et al., 1984). tissues (Honn et al., 1981;Bennett et al., 1982;Porteder et 3H-PG (nM) al., 1984). However, the type of PG produced by tumour 0.12 -b cells as well as the capacity of the tumour cells to synthesise PGs are highly variable (Malachi et al., 1981;Ziboh et al., 0.10 \1981). Cummings and Robertson (1977)  3H-PGE1 with a Kd value of about 4 nM for 3H-PGE1 binding (0) 3H-PGE,-binding; 3H-PGE2-binding (A). and of 6 nM 3H-iloprost binding on the average. While for Bm.: binding capacity in fmol mg-' protein; Kd: (dissociation constant) in nM; not investigated. ap < 0.05, bp < 0.0 , Cp < 0.001, vs the control group (without irradiation). dp <0.05, vs the irradiated control group. Significant (P<0.05) increase in 3H-iloprost receptors expressed on hypernephroma cells was found after pre-treatment with ASA which also led to a significant (P < 0.05) increase in the binding affinity. Irradiation caused significant (P<0.001) depression of 3H-iloprost receptors while the binding affinity remained unchanged.
Irradiated ASA-pretreated cells had a similar binding capacity to the plane control group and a significantly (P < 0.05) higher one as compared with the irradiated control group. most cells 3H-PGE, as well as 3H-iloprost seem to bind to two binding classes with different affinity (Virgolini et al., 1992), our results on cultured human hypernephroma cells indicate a single class of high affinity sites for either ligand. Furthermore, the competition studies show that 3H-iloprost could recognise 3H-PGE, sites as does 3H-PGE, with 3Hiloprost sites vice versa, pointing to a common PGEI/PGE2 binding site on human hypernephroma cells. As with nonmalignant cells (Virgolini et al., 1992) 3H-PGE2 did not bind to PGE1/PGE2 binding sites suggesting a distinct class of 3H-PGE2 binding sites.
In order to study the effects of ASA on the PG-system of human hypernephroma cells an incubation was performed for 24 h. ASA significantly depressed PG-production and increased PG receptor densities and binding affinities, whereas cellular viability remained unchanged. These results indicate that ASA renders human hypernephroma cells more sensitive for binding PGE,, PGE2 as well as iloprost. Whether or not there exists a down-regulation mechanism on the basis of an increased PG production by human hypernephroma cells is not yet clear.
Recently a number of studies in murine tumours have shown that indomethacin significantly increases radioresponse of tumours which produce PGs (Furuta et al., 1988a(Furuta et al., , 1988b. Initially it was proposed that the augmentation of radioresponse induced by indomethacin could be due to lowering PGs in tumour tissues. The effects of PGs on radioprotection are, however, not clear: some authors reported that PGs might act as radioprotectors (Hanson & Ainsworth, 1985;Walden et al., 1987), whereas others (Milas et al., 1990) proposed an opposite effect.
PGs exert their effects through interaction with specific cell surface receptors. There is little information about PG receptor expressed by tumour cells after radiotherapy. We were now able to show that although the number of PG binding sites was significantly diminished after irradiation with 'Co, no significant change in the number of PG binding sites was found for human hypernephroma cells after pre-incubation with ASA. The results indicate that the PG synthesis inhibitor ASA could protect against an irradiation effect on PG receptors, modifying the radioresponse of human hypernephroma cells. Bmax: binding capacity in fmol mg-' protein; Kd: (dissociation constant) in nM; -: not investigated.
ap<0.05, bP<0.01, CP<0.001, vs the control group (without irradiation). dP<0.01, vs the irradiated control group. Significant (P < 0.01) increase in 3H-PGE, receptors expressed on human hypernephroma cells was found after pre-treatment with ASA which also led to a significant (P < 0.05) increase in the binding affinity. Irradiation caused a significant (P < 0.001) depression of 3H-PGE, receptors without change in the binding affinity. Irradiated ASA-pre-treated cells had a similar binding capacity to the plane control group and a significant (P<0.01) higher one as compared with the irradiated control group. 0.01, vs the control group (without irradiation). CP < 0.05, vs the irradiated control group. Significant (P < 0.05) increase in 3H-PGE2 receptors expressed on human hypernephroma cells was found after pre-treatment with ASA which also led to a significant (P < 0.05) increase in the binding affinity. Irradiation caused a significant (P < 0.01) depression of 3H-PGE2 receptors without change in the binding affinity. Irradiated ASA-pretreated cells had a significantly (P < 0.05) higher number of 3H-PGE2 receptors as irradiated ASA-untreated cells, however values were still significantly (P<0.05) lower than in the control group.
a  The presence of PGs as well as PG ri human hypernephroma cells suggests . tumour growth and possibly its metastas already been reported that certain syndro hypernephromas such as the osteolytic secutive hypercalcaemia might result fror hypernephroma cells (Robertson et al., Robertson, 1977). Furthermore, increases tumours has been associated with a inhibitors of PG synthesis including ASA were reported to retard tumour growth mation (Lynch et al., 1978;Bennett,, 19 cAMP is a second messenger mediating the actions of PGs and also connected with normal and malignant cell proliferation. It is likely that changes in intracellular cAMP-1.0 10 100 production due to PG stimulation may greatly influence tumour cell growth and even metastases formation (Heidrick & Ryan, 1971;Sheppard, 1972). However, both reduced ation on cAMP-pro cAMP levels (Malachi et al., 1981;Goldberg et al., 1975; by different concen- Stevens et al., 1978) and elevated cAMP contents (Minton et  tissues. This may correlate to the different PG levels found Dup irradiated; -0-: for various tumour tissues. For cultured human hyperneph-P < 0.01 vs the basal roma cells we found no significant difference in basal cAMP levels in both non-ASA-treated and ASA-treated cells. However, the PG-stimulated cAMP-production was significantly higher in the ASA-treated group as compared with the control group. This suggests that down-regulation of PG recepeceptors in cultured tors was associated with decreased adenylate cyclase activity a role for PGs in stimulated by PGs.

,es formation. It has
We here show that ASA modifies the radioresponse of mes of patients with human hypernephroma cells by preventing the irradiationprocess and con-induced decrease of PG receptors, which might cause a m PGs produced by corresponding decrease in PG-induced membrane adenylate 1975; Cummings & cyclase activity. I PG production by Lggression, whereas k and indomethacin and metastases for-182).
The authors are grateful for the expert technical assistance of Ingrid Teufel and Regina Haslinger. This study was supported by a grant of the 'Kommission fur Onkologie', University of Vienna, Vienna, Austria.