Influence of single and multiple doses of amifostine on the efficacy and the pharmacokinetics of carboplatin in mice.

We have previously reported that amifostine potentiates the anti-tumour activity of carboplatin in mice. The present study was carried out in well-established human ovarian cancer xenografts OVCAR-3, A2780 and FMa grown subcutaneously in the nude mouse. It was found that a single dose of amifostine resulted in a higher increase in the anti-tumour activity of carboplatin than three doses of amifostine. A single dose of amifostine increased the AUC (area under the curve) values of total platinum in plasma ultrafiltrate (30.1 vs 18.2 microM x h), liver (307.7 vs 236.4 nmol g(-1) x h), kidney (500.8 vs 368.3 nmol g(-1) x h) and OVCAR-3 tumour tissue (184.0 vs 146.8 nmol g(-1) x h). Despite this increase in total platinum, a decrease in platinum (Pt)-DNA adduct levels was observed in liver, kidney and bone marrow, which was significant in liver. In tumour tissue an insignificant increase in Pt-DNA adduct levels, specifically the Pt-GG adduct, was observed after treatment with a single dose of amifostine, which may explain the increase in anti-tumour activity. The increase in the AUC of total platinum was probably caused by a reduction in body temperature, which was most severe after three doses of amifostine. The extreme hypothermia may be the reason that three doses of amifostine resulted in less potentiation of the efficacy of carboplatin.

Carboplatin [cis-diammine( 1,1 -cyclobutanedicarboxylato)platinum(II)] was developed as a second-generation platinum compound with less nephrotoxicity than cisplatin. Its anti-tumour activity is assumed to result from the formation of platinum (Pt)-DNA adducts. The clinical use of carboplatin is limited by myelosuppression at a dose in the steep part of the dose-response curve. Therefore, much effort has been put into reducing the toxic side-effects to allow the administration of higher doses of carboplatin.
Amifostine [Ethyol, WR-2721, S-2-(3-aminopropylamino)ethylphosphorothioic acid], initially developed as a radioprotector, is approved for use as a protector against chemotherapy-induced toxicities in the USA and Europe (van der Vijgh and Peters, 1994). A selective protection against the side-effects of platinum compounds has been observed in clinical and preclinical studies (Treskes et al, 1992a(Treskes et al, , 1994Capizzi, 1994;van der Vijgh and Peters, 1994). Amifostine is the prodrug of the aminothiol compound WR-1065 (Figure 1), which inhibits DNA platination (Treskes et al, 1992b). The selective protection is based on the preferential formation and uptake of this active metabolite in nontumour tissues (Yuhas, 1980;Brown et al, 1988;Calabro-Jones et al, 1988;Shaw et al, 1988).
In vitro experiments have shown that the modifying action of amifostine was protection rather than rescue from toxicity (Treskes et al, 1992b). Considering these results and the rapid uptake and clearance of WR-1065 by non-tumour tissues after the (i.p. or i.v.) administration of amifostine (Utley et al, 1984;Shaw et al, 1988Shaw et al, , 1994, optimal protection would be achieved when amifostine is administered shortly before the platinum drug. Efficient protection against cisplatin-induced nephrotoxicity was observed in mice when amifostine was administered 5 min or 30 min before cisplatin (Treskes et al, 1992a). Protection against carboplatin-induced myelotoxicity was more obvious when amifostine was given 5 min instead of 30 min before carboplatin (Treskes et al, 1994). With respect to the long elimination half-life of non-protein-bound carboplatin in comparison with amifostine (van der Vijgh, 1991), greater myeloprotection might be achieved by multiple doses of amifostine when combined with carboplatin. Such an approach is incorporated into clinical trials in which amifostine is given three times, just before and 2 and 4 h after carboplatin (Betticher et al, 1995;Vermorken et al, 1995). To date, clinical data are available for only a small number of patients. The results do suggest that amifostine can reduce the duration of thrombocytopenia and hospitalization (Betticher et al, 1995).
The use of modulating agents can only be successful when these compounds do not interfere with the anti-tumour activity of the cytotoxic agent. Studies in tumour-bearing nude mice have demonstrated that amifostine administered 5 min before the platinum compound did not reduce the anti-tumour efficacy of cisplatin (Treskes et al, 1992a) and carboplatin (Treskes et al, 1994). In the case of carboplatin even a potentiation of the antitumour activity was noticed when amifostine was given once, whereas amifostine itself did not affect tumour growth (Treskes et al, 1994). However, the influence of multiple doses of amifostine on the anti-tumour activity of carboplatin was not studied.
In the present experiments we investigated the influence of multiple doses of amifostine on the efficacy of carboplatin in tumour-bearing nude mice. We also investigated the influence of one and three doses of amifostine on the pharmacokinetics of carboplatin and the formation of Pt-DNA adducts in order to interpret the efficacy of carboplatin in the presence of amifostine. Furthermore, we determined the influence of a single dose and multiple doses of amifostine on the body temperature in nude mice.

MATERIALS AND METHODS Chemicals
Paraplatin (150 mg of lyophilized carboplatin and 150 mg of mannitol) was obtained from Bristol-Myers Squibb (Woerden, The Netherlands) and reconstituted with 15 ml of glucose 5% before use. Amifostine (500 mg of lyophilized WR-2721 and 500 mg of mannitol) was obtained from US Bioscience (West Conshohocken, PA, USA) and reconstituted with 9.3 ml of sterile water. Before use the compound was further diluted with 0.9% sodium chloride.
Xenografts and treatment schedules Female athymic nude mice (Harlan/Cpb, Zeist, The Netherlands), housed and fed as described previously (Boven et al, 1985), were inoculated subcutaneously (s.c.) in both flanks with fragments (2-3 mm in diameter) of the human ovarian cancer xenografts OVCAR-3, A2780 or FMa. OVCAR-3 is a poorly differentiated serous adenocarcinoma, A2780 an undifferentiated carcinoma and FMa a poorly differentiated mucinous adenocarcinoma.
Treatment was started when tumours had reached approximately 300 mm3 in size in the pharmacokinetic study and 50-150 mm3 in the anti-tumour activity studies (designated as day 0). Amifostine injections were given i.p., whereas carboplatin was administered in a tail vein. Drug doses were derived from previous experiments in which carboplatin 60 mg kg-' given on days 0 and 7 was considered the maximum tolerated dose (MTD) allowing 10% weight loss within 2 weeks after the first injection (Treskes et al, 1994). One injection of amifostine (200 mg kg-') given 5 min before carboplatin enhanced the anti-tumour activity in the OVCAR-3 xenograft model. In those experiments the MTD of carboplatin preceded by one injection of 200 mg kg-' amifostine could be increased to 90 mg kg-'. In the current study, mice were treated with 60 mg kg-' carboplatin alone or in combination with a single or three doses of 200 mg kg-' amifostine to investigate the influence of amifostine on the anti-tumour activity and pharmacokinetics of carboplatin. Amifostine was administered 5 min before carboplatin as a single dose or 5 min before and 2 and 4 h after carboplatin when administered three times. The 90 mg kg-' dose of carboplatin was also investigated for its anti-tumour activity when administered under the protection of three doses of amifostine.

Anti-tumour activity
To investigate the influence of three doses of amifostine on the anti-tumour activity of carboplatin, mice bearing bilateral OVCAR-3, A2780 or FMa xenografts (5-6 mice per group) were treated on days 0 and 7 with 60 mg kg-' carboplatin alone or in combination with three times 200 mg kg-' amifostine. A control group (untreated) and a group treated with three doses of amifostine alone were included. In OVCAR-3and FMa-bearing mice an additional group was treated on days 0 and 7 with 90 mg kg-' carboplatin in combination with three doses of amifostine. In a second experiment the influence of one vs three doses of amifostine on the anti-tumour activity of carboplatin 60 mg kg-' was investigated in mice bearing OVCAR-3 or FMa xenografts, treated on days 0 and 7. In this experiment, additional groups of mice were control (untreated) or treated with carboplatin alone.
In the first experiment mice were weighed twice a week and in the second once (OVCAR-3 and FMa xenografts) or twice (A2780 xenografts) a week. Tumours were measured weekly with a slide caliper by the same observer. Tumour volume was calculated as length x width x height x 0.5 in mm3 and expressed relative to the volume at the start of the treatment. The anti-tumour activity was expressed as the percentage of growth inhibition [1 -(mean of the relative volumes of the treated tumours divided by that of the control tumours) x 100%]. In principle this was calculated on the day of its maximum value, reached within 5 weeks after the last day of treatment. In addition, the number of days for a tumour to reach four times its volume from that at the start of treatment (TDIO4), was calculated.

Pharmacokinetics
Mice bearing OVCAR-3 xenografts were treated with 60 mg kg-' carboplatin alone or in combination with 1 x or 3 x 200 mg kg-' amifostine. Of each group, three mice per time point were bled from the axillary vein under ether anaesthesia at 0.5, 1, 1.7, 3, 5, 8, 12 or 24 h after carboplatin administration. Thereafter, liver, kidney and tumours were removed. Bone marrow was collected by flushing both femurs with RPMI medium and pooled from the three mice per time point of each group. The plasma samples were ultrafiltrated by MPS-1 systems provided with YMT filters (Amicon, Capelle a/d IJssel, The Netherlands). Plasma ultrafiltrate and parts of liver, kidney and tumours were stored at -20°C until analysis of total platinum. Bone marrow and parts of liver, kidney and tumours were stored at -80°C until analysis of Pt-DNA adducts.
For platinum analysis the samples were pretreated as follows. Plasma ultraflltrate samples were diluted 1:1 with 0.15 M sodium chloride/0.4 M hydrochloric acid before measurement of the platinum concentration. Tissue samples of 100 mg were digested in 1.0 ml of concentrated nitric acid in a Teflon bomb at 170°C for 2 h. After cooling, the samples were transferred to glass tubes and evaporated under nitrogen after addition of 20 gl of 1.7 M sodium chloride. The residues were reconstituted in 250 pl of 0.15 M sodium chloride/0.2 M hydrochloric acid. Platinum concentrations were analysed by flameless atomic absorption spectrophotometry using a Spectra AA-300 Zeeman AAS (Varian, Houten, The Table 1 Efficacy and toxicity of carboplatin ± amifostine in well-established s.c. human ovarian cancer xenografts in female nude mice (n 2 5 per group) treated weekly x 2 with 3 x 200 mg kg-' amifostine, 60 mg kg-' carboplatin, 60 mg kg-' carboplatin + 3 x 200 mg kg-' amifostine or 90 mg kg-' carboplatin + 3 x 200 mg kg-'  (0) Netherlands). Standards of blank plasma ultrafiltrate and tissue spiked with carboplatin were treated in the same way as the samples.
To determine Pt-DNA adducts, tissue samples were ground to cell suspensions in 2 ml of Tris-EDTA supplemented with 0.5 M ammonium bicarbonate. DNA was isolated as described elsewhere (Blommaert et al, 1995). From the liver samples the cell nuclei were isolated before DNA isolation according to a previously described method (Roggeband et al, 1993). After digestion the isolated DNA samples were chromatographed on a Mono Q anion-exchange column (Pharmacia, Woerden, The Netherlands) and appropriate column fractions were analysed in a competitive enzyme-linked immunosorbent assay (ELISA) (Fichtinger-Schepman et al, 1987).
With this procedure four platinum-containing (di)nucleotides could be identified (Fichtinger-Schepman et al, 1987;Blommaert et al, 1995): Pt-GG and Pt-AG (bifunctional adducts of carboplatin with two adjacent guanines or with adenine adjacent to guanine), G-Pt-G [bifunctional adduct of carboplatin with two non-adjacent guanines either in the same strand (intrastrand crosslink) or with two guanines in the opposite DNA strands (interstrand crosslink)] and Pt-G (carboplatin monofunctionally bound to a guanine residue).
Area under the curve (AUC) values of the concentration-time curves of total platinum and of Pt-DNA adducts were calculated from the mean concentrations at each time point, over 0.5-24 h after the carboplatin administration, using the trapezoidal rule.

Temperature
Non-tumour-bearing nude mice were either injected with 60 mg kg-' carboplatin alone (n = 3) or in combination with 200 mg kg-' amifostine, administered once (n = 3) or three times (n = 3). Temperature changes were monitored for 24 h by measuring body temperature intrarectally by a Lameris Ellab-Instruments thermocouple (Lameris, Utrecht, The Netherlands).

Statistics
Results of the pharmacokinetic and the anti-tumour activity studies were evaluated with Student's t-test.

Anti-tumour activity
The anti-tumour activity of carboplatin in the three human ovarian cancer xenografts was expressed as the percentage of growth inhibition and the TDI<4 (Table 1). Carboplatin at the MTD of 60 mg kg-' was active in OVCAR-3 and FMa xenografts, whereas no significant influence on the growth of A2780 tumours was Table 2 Efficacy and toxicity of carboplatin ± amifostine in well-established s.c. xenografts OVCAR-3 and FMa in female nude mice (six per group) treated weekly x 2 with 60 mg kg-' carboplatin alone or in combination with 1 x 200 mg kg-' amifostine or 3 x 200 mg kg-' amifostine Lo.
. ). I P < 0.05 compared with treatment C; 2p < 0.05 compared with treatment C + IA 90 mg kg-' carboplatin and three doses of amifostine was the most toxic (maximum weight loss 15.9% and 20.3% respectively), followed by 60 mg kg-' carboplatin plus three doses of amifostine (9.5% and 13.5% respectively). Amifostine given three times induced more weight loss than when given only once (latter data not shown). Except for the carboplatin 90 mg kg-1 plus three doses of amifostine schedule, mice had recovered from weight loss on day 14 of the treatment experiment. In Table 2 it is shown that recovery on day 14 was better in mice receiving carboplatin plus one dose of amifostine than in animals receiving three doses of amifostine. Toxic deaths were observed in mice receiving schedules that contained three doses of amifostine, even without carboplatin. Animals died within 4 days after the injections and this was preceded by excessive weight loss. Figure 4 AUC values of Pt-DNA adducts in liver, kidney, OVCAR-3 tumour tissue and bone marrow from 0.5 to 24 h after treatment with carboplatin alone (C) or in combination with 1 x 200 mg kg-' amifostine (C + 1A) or 3 x 200 mg kg-' amifostine (C + 3A). 'P< 0.05 compared with treatment C; 2p< 0.005 compared with treatment C + 1 A observed. In OVCAR-3 and FMa xenografts the administration of a 1.5-fold increased dose of carboplatin (90 mg kg-'), which can only be administered under the protection of amifostine (Treskes et al, 1994), was significantly more effective than 60 mg kg-' carboplatin alone or combined with amifostine (P < 0.01). Amifostine, given as a single dose (Treskes et al, 1994) or given three times, did not have any anti-tumour activity. When 60 mg kg-' carboplatin was combined with three doses of amifostine no significant change in growth inhibition was observed when compared with 60 mg kg-' carboplatin alone. Our results on carboplatin and three doses of amifostine did not correspond with the earlier reported significant potentiation of the anti-tumour activity of carboplatin by one dose of amifostine (Treskes et al, 1994). A second experiment was performed in which we investigated the presence of a doseor a scheduledependent interaction between amifostine and carboplatin. Therefore, we examined the influence of one vs three doses of amifostine on the anti-tumour activity of carboplatin in OVCAR-3and FMa-bearing nude mice (Figure 2, Table 2). When carboplatin was combined with three doses of amifostine a slight increase in the anti-tumour activity was observed in OVCAR-3-bearing mice when compared with carboplatin alone (TDI,4 38.8 vs 33.5 days, P < 0.02). However, when only one dose of amifostine was administered a more pronounced potentiation of the anti-tumour activity of carboplatin was observed in OVCAR-3-bearing mice when compared with carboplatin alone (growth inhibition 92.7% vs 88.5% and TDI 4 43.7 vs 33.5 days, P < 0.001), which was comparable to that described before (Treskes et al, 1994). In FMa-bearing mice the same trend was observed, although the difference was not significant (growth inhibition 82.0% vs 76.0% and TD}14 36.8 vs 32.4 days).
As a measure for toxicity, maximum weight loss and the body weight on day 14, both expressed as a percentage of the weight at the start of treatment, as well as the number of toxic deaths have been summarized in Tables and 2. Nadir weight loss after treatment with carboplatin and amifostine was observed 3-4 days after treatment (Treskes et al, 1994). In Table 2 maximum weight loss was not documented, because the weight was measured weekly. In OVCAR-3-and FMa-bearing nude mice the combination of Pharmacokinetics Concentration-time curves of platinum in plasma ultrafiltrate and tissues from OVCAR-3-bearing nude mice after treatment with carboplatin alone or in combination with one or three doses of amifostine are shown in Figure 3. In plasma ultrafiltrate as well as in normal and tumour tissues, platinum concentrations were higher after treatment with carboplatin and amifostine than with carboplatin alone, resulting in an increase in AUC values as shown in Figure 3. In plasma ultrafiltrate, increases of approximately 1.7fold and 2-fold were observed when compared with carboplatin alone, of which only the 2-fold increase after three doses of amifostine was statistically significant (P < 0.05). In tissues a significant increase of approximately 1.3-fold and 1.5-fold (P < 0.05) was observed after treatment with a single and three doses of amifostine, respectively. The difference between the AUC values after the singleand the three-dose schedule was only significant in liver (P < 0.05).
The AUC values of Pt-DNA adduct levels are given in Figure 4. Each column shows the AUC value of the total DNA platination, i.e. the sum of the four types of adducts determined in the digested DNA samples. In the normal tissues the total Pt-DNA adduct levels were lower after treatment with a single dose of amifostine than treatment with carboplatin alone. Decreases of approximately 0.8-fold were observed. A significant decrease was only observed in the liver (P < 0.05), mainly as a result of decrease in the G-Pt-G adducts. In tumour tissue a slight, insignificant increase of approximately 1.1-fold was seen. When comparing these results with the Pt-DNA adduct levels after treatment with carboplatin and three doses of amifostine, a 1.2to 1.6-fold increase in total Pt-DNA adducts was observed in normal tissues, which was mainly as a result of higher levels of Pt-G and G-Pt-G adducts. The differences in total Pt-DNA adduct levels between the singleand the three-dose schedules were only statistically significant for the liver samples (P < 0.005), whereas significant differences in G-Pt-G levels were found in liver (P < 0.005), kidney (P < 0.01) and tumour tissue (P <0.05) and in Pt-G levels in liver and kidney (P < 0.05).
Because the Pt-DNA adduct levels were not increased proportionally to the increase in total platinum, we calculated the ratio AUC Pt-DNA adducts/AUCtotal platinum' In liver tissue this ratio was 0.23 after treatment with carboplatin alone, 0.13 after carboplatin plus one dose of amifostine and 0.18 after carboplatin plus three doses of amifostine. In kidney these ratios were 0.51, 0.31 and 0.40 respectively. In tumour tissue ratios of 0.56, 0.48 and 0.46 were observed for the three treatment schedules. when carboplatin was combined with a single dose of amifostine 00°¢:M__ (Treskes et al, 1994). Our second experiment confirmed that a poO single dose of amifostine resulted in a significant potentiation of the anti-tumour activity of carboplatin, whereas treatment with 0 W three doses of amifostine resulted in a less pronounced increase in the efficacy of carboplatin (Table 2, Figure 2).

K)_.
From our results on the potentiation of the efficacy of carboplatin by amifostine, we anticipated the presence of a pharmacological drug interaction. In plasma ultrafiltrate, kidney, liver and tumour tissue the platinum concentrations were raised significantly after treatment with amifostine. This pharmacokinetic interaction seemed to be dependent on the treatment schedule. Three In the normal tissues a single dose of amifostine resulted in a small decrease in the Pt-DNA adduct levels, mainly due to the an values of the body temperatures of non-tumour-bearing level of G-Pt-G adducts, despite the increase in total platinum. lice after treatment with carboplatin alone, carboplatin in Although the observed decrease was significant only in liver, these ation with a single dose of amifostine as well as with three results might suggest a protection by amifostine as was observed f amifostine are summarized in Figure 5. Carboplatin alone in mice (Treskes et al, 1994). After treatment with three doses of decrease the body temperature of 370C in the 24 h amifostine higher Pt-DNA adduct levels were observed in normal ng injection. After a single dose of amifostine the lowest tissues than with a single dose of amifostine, which was most iture was 34.50C, which was observed I h after administraprobably due to the higher total platinum levels. These data might lice rapidly recovered within 6 h. After three doses of indicate that, when amifostine protects against platinum-induced ine, however, a mean body temperature of 25°C was toxicities by reduction of Pt-DNA adduct levels, this protection d 7 h after the initiation of treatment, which only started to will be less pronounced after three doses than after a single dose of 22 h after the first injection.
amifostine. However, this has never been confirmed by toxicity studies, because only the influence of a single dose of amifostine JSSION on the carboplatin-induced toxicities has been investigated (Treskes et al, 1994). The small increase in Pt-DNA adduct levels tine is an agent that protects against cisplatin-induced in normal tissues after treatment with three doses of amifostine oxicity and carboplatin-induced myelotoxicity without was not proportional to the rise of the total platinum concentrag the anti-tumour activity of these platinum compounds in tions, as calculated by the ratio AUC PtDNA adducts/AUCtotal pl.tinum' The ental human tumours (Treskes et al, ]992a, 1994). The ratios after treatment with carboplatin and amifostine were of interference with the anti-tumour activity had already reduced when compared with the ratio after treatment with carbopected from the relatively low reaction rates of carboplatin platin alone. This relative decrease in Pt-DNA adduct levels was ifostine and its main metabolites (Treskes et al, 1991)  if amifostine with nitrogen mustard (Valeriote and Tolen, adducts was observed in tumour tissue after treatment with one as nd that with melphalan (Millar et al, 1982). In 25 patients well as with three doses of amifostine. Upon calculating the ratio n-small-cell lung cancer a high response rate of 64% has AUC Pt-DNA adducts/AUC Otal platinun' a smaller decrease was observed for ported upon treatment with cisplatin and vinblastine in carboplatin combined with one dose of amifostine than when ition with amifostine (Schiller et al, 1996), whereas compared with the ratios in normal tissues. This might indicate z rates of 25-30% are recorded after standard chemothat the protective effect of amifostine was mainly restricted to the In another study in 21 non-small-cell lung cancer patients, normal tissues. This is in agreement with the observed selective it with carboplatin and amifostine was at least as active as uptake of WR-1065 in normal tissues when compared with the it with carboplatin alone (Betticher et al, 1995). uptake in tumour tissues (Shaw et al, 1994). As a consequence, no influence of amifostine on the anti-tumour activity of protection of the tumour was to be expected in the efficacy studies, tin seemed to be doseor schedule-dependent in our which was indeed the case. Moreover, we even found a potentiaents. When carboplatin was combined with three doses of tion of the anti-tumour activity, especially in the case of one dose ne no interference with the anti-tumour activity was of amifostine. Unfortunately, the difference in potentiation J ( Furthermore, it is questionable whether all four Pt-DNA adducts are responsible for the anti-tumour activity. When the extent of anti-tumour activity was compared with the value of each of the four types of adducts then a positive relation with the major adduct Pt-GG could be established. It is not clear whether amifostine had an influence on the adduct formation or removal when comparing the levels at the individual time points. Owing to the rather high standard deviations differences between the formation and removal of Pt-DNA adducts could not be clearly distinguished. The reason for the observed amifostine-carboplatin pharmacokinetic interaction and the difference between one and three doses of amifostine added to carboplatin might be the observed amifostine-induced hypothermia. The extreme reduction in body temperature after treatment with three doses of amifostine will result in a peripheral vasoconstriction affecting the renal clearance of carboplatin. This might induce an increase in the AUC values as observed in this study. Although hypothermia leads to a reduction in the cytotoxicity of cisplatin in vitro (Page et al, 1987), no tumour protection was observed in our in vivo experiments. The extreme hypothermic conditions induced by 3 x 200 mg kg-' amifostine, however, did result in a less pronounced potentiation of the efficacy of carboplatin. Because hypothermia is related to the dose of amifostine, it also explains the suggested tumour protection by the compound when given as a single dose of 400 mg kg-' (Twentyman, 1983). From other studies in which reduction in body temperature by amifostine has been described in Balb/c ( Van der Wilt et al, 1992) and severe combined immunodeficient (SCID) mice (Paine et al, 1996), it appears that the extent of this side-effect varies between different strains of mice. The extreme hypothermia in our nude mice is most probably the reason for the observed toxic deaths (Table 1). The extent of hypothermia in mice is not representative of the human situation because no clear decrease in body temperature has been observed in patients. Therefore, the possible influence of amifostine on the anti-tumour activity of carboplatin still needs to be investigated in patients.
In conclusion, amifostine does not reduce the anti-tumour activity of carboplatin in tumour-bearing nude mice. When given once it even potentiates the anti-tumour activity of carboplatin. Amifostine has an influence on the pharmacokinetics of carboplatin, most probably explained by altered drug distribution caused by hypothermia, resulting in higher platinum concentrations in normal and tumour tissues. The relative reduction in total Pt-DNA adduct levels in these tissues suggests a selective protection of the normal tissues, but not of the tumour tissue. Three doses of amifostine resulted in an extreme reduction of the body temperature as well as toxic deaths and a less pronounced potentiation of the anti-tumour activity. Extrapolating these data to the clinical application of amifostine is difficult because the amifostineinduced hypothermia was not observed in patients. However, our data confirm the suggested selective protection of normal tissues by amifostine, which makes amifostine a promising modulating agent. van der Vijgh WJF and Peters GJ (1994) Protection of normal tissues from the cytotoxic effects of chemotherapy and radiation by amifostine (Ethyol): Preclinical aspects. Semin Oncol 21: 2-7 Van Der Wilt CL, Van Laar JAM, Gyergyay F, Smid K and Peters GJ (1992) Biochemical modification of the toxicity and antitumor effect of 5-fluorouracil and cisplatinum by WR-2721 in mice. Eur J Cancer 28A: 2017-2024