Effect of MEN 10755, a new disaccharide analogue of doxorubicin, on sarcoplasmic reticulum Ca²⁺ handling and contractile function in rat heart

Abstract

1. The use of anthraquinone antineoplastic agents is limited by their cardiac toxicity, which is largely due to activation of the sarcoplasmic reticulum (SR) Ca²⁺ release channel (ryanodine receptor). MEN 10755 is a new disaccharide analogue of doxorubicin. We have evaluated its effects on SR function and its toxicity in isolated working rat hearts.

2. In rat SR vesicles, doxorubicin stimulated [³H]-ryanodine binding by increasing its Ca²⁺-sensitivity. At 1 μM Ca²⁺, ryanodine binding increased by 15.3±2.5 fold, with EC₅₀=20.6 μM. Epirubicin produced a similar effect, i.e. 9.7±0.6 fold stimulation with EC₅₀=11.1 μM. MEN 10755 increased ryanodine binding by 1.9±0.3 fold (P<0.01 vs doxorubicin and epirubicin), with EC₅₀=38.9 μM.

3. Ca²⁺-induced Ca²⁺ release experiments were performed by quick filtration technique, after SR loading with ⁴⁵Ca²⁺. At 2 μM Ca²⁺, doxorubicin (50 μM) increased the rate constant of Ca²⁺ release to 82±5 s⁻¹ vs a control value of 22±2 s⁻¹ (P<0.01), whereas 50 μM MEN 10755 did not produce any significant effect (24±3 s⁻¹).

4. Ca²⁺-ATPase activity and ⁴⁵Ca²⁺-uptake were not significantly affected by doxorubicin, its 13-dihydro-derivative, epirubicin, MEN 10755 and the 13-dihydro-derivative of MEN 10755, at concentrations ⩽100 μM.

5. In isolated heart experiments, administration of 30 μM doxorubicin or epirubicin caused serious contractile impairment, whereas 30 μM MEN 10755 produced only minor effects.

6. In conclusion, in acute experiments MEN 10755 was much less cardiotoxic than equimolar doxorubicin or epirubicin. This result might be accounted for by reduced activation of SR Ca²⁺ release.

Introduction

Anthraquinones such as doxorubicin and epirubicin are widely used as antineoplastic agents, but their clinical administration is often limited by side effects, and particularly by cardiac toxicity. Acute cardiac effects include a transient positive inotropic response followed by sustained negative inotropic action, prolonged time to peak twitch tension, and decreased relaxation rate, whereas chronic doxorubicin toxicity is characterized by the progressive development of cardiomyopathy, leading to congestive heart failure (Buzdar et al., 1985; Jensen, 1986; Hagane et al., 1988; Doroshow, 1991; Rhoden et al., 1993; Wang & Korth, 1995; Nysom et al., 1998).

Several molecular mechanisms might contribute to these toxic effects. Doxorubicin stimulates free radical production, interferes with mitochondrial respiration, and affects sarcoplasmic reticulum (SR) function. According to a widely held opinion, cardiac toxicity is largely due to interference with intracellular Ca²⁺ homeostasis (Jiang et al., 1994; Temma et al., 1994; Halili-Rutman et al., 1997; Maeda et al., 1998; Feng et al., 1999), and particularly to stimulation of SR Ca²⁺ release. Doxorubicin binds to the SR Ca²⁺ channel (also known as ryanodine receptor) and favours channel opening, leading to increased cytosolic Ca²⁺ (Zorzato et al., 1985; Abramson et al., 1988; Kim et al., 1989; Ondrias et al., 1990; Pessah et al., 1990; Tian et al., 1991; Boucek et al., 1993). Cytosolic Ca²⁺ overload transiently stimulates muscle contraction, but triggers several injurious mechanisms, such as activation of Ca²⁺-dependent proteases and phospholipases, finally leading to contractile dysfunction and cellular necrosis (Nayler, 1988).

Strong efforts have been performed to produce doxorubicin derivatives that retain antineoplastic efficacy while showing reduced cardiac toxicity. Recently, disaccharide analogues of doxorubicin have been synthesized, in which the amino sugar moiety appears as the second residue bound in axial orientation to the first residue, that in turn is characterized by the substitution of the amino group for a hydroxyl group (Arcamone et al., 1997). One of these analogues, MEN 10755 (7-O-[2,6-dideoxy-4-O-(2,3,6-trideoxy-3-amino-α-L-lyxo-hexopyranosyl) -α-  L - lyxo-hexopyranosyl]-4-demethoxy-14-hydroxydaunomycinone hydrochloride), has proved to be superior to doxorubicin in the treatment of human tumour xenografts (Pratesi et al., 1998). Before MEN 10755 can be proposed for clinical use, it is necessary to assess its side effects, and particularly its cardiac effects. In this work we have compared the effects of MEN 10755, doxorubicin, and epirubicin (the 3′-epimer of doxorubicin) on cardiac ryanodine receptors. The 13-dihydro derivatives of doxorubicin and MEN 10755, which are formed in vivo during anthraquinone catabolism, have also been tested. Finally, we have compared the acute toxicity of MEN 10755, doxorubicin, and epirubicin in the isolated perfused rat heart.

Methods

Preparation of cellular fractions

Male Wistar rats (275–300 g body weight), fed with standard diet, were anaesthetized with a mixture of ether and air. After injection of 1000 u sodium heparin in the femoral vein, the heart was quickly excised and the ventricles were finely minced and homogenized in five volumes of 300 mM sucrose and 10 mM imidazole (pH 7.0 at 4°C) by 15+15 passes in a Potter-Elvejheim homogenizer set at 800 r.p.m. and kept in a cold room at 4°C. The homogenate was then used to prepare a microsomal fraction enriched in SR, as described in detail elsewhere (Zucchi et al., 1995a). The protein content of each fraction was determined by the Lowry method, using bovine serum albumin as a standard.

Assay of [³H]-ryanodine binding

High affinity ryanodine binding was assayed at different free Ca²⁺ concentrations, as described previously (Zucchi et al., 1994; 1995a). Briefly, vesicles were incubated at 37°C in a buffer containing imidazole 25 mM (pH 7.4 at 37 °C), 1 M KCl, 0.2 to 50 nM [³H]-ryanodine (6 Ci mmol⁻¹), EGTA 0.95 mM, and variable amounts of CaCl₂, in order to obtain the desired free Ca²⁺ concentration. After 60 min, the binding reaction was stopped by filtration through cellulose nitrate filters with pores of 0.45 μM (Sartorius, Göttingen, Germany), presoaked in 25 mM imidazole and 1 M KCl (washing buffer). The filters were then washed with 2×5 ml aliquots of washing buffer and shaken overnight in 8 ml of scintillation fluid (Optiphase II, LKB, Turku, Finland). Radioactivity was counted at 50% efficiency in an LKB Wallac 1214 scintillation counter (Turku, Finland). Incubations were performed in duplicate, and non-specific binding was measured in the presence of 10 μM unlabelled ryanodine. The difference between the counts of duplicate samples was <10% in all cases.

Assay of SR Ca²⁺ release

SR Ca²⁺-induced Ca²⁺ release was determined as described previously (Zucchi et al., 1995b), with the only difference that vesicle loading was performed in a medium containing 100 μM ⁴⁵CaCl₂, as described by Chu et al. (1988). ⁴⁵Ca release was induced by washing the loaded vesicles with a release buffer containing HEPES-potassium 20 mM (pH 6.8), KCl 100 mM, and various amounts of CaCl₂, buffered with EGTA (1 mM) to obtain the desired free Ca²⁺ concentration. A rapid filtration system with time resolution on the order of 10 ms was used (Dupont, 1984; Moutin & Dupont, 1988), and the rate constant of quick Ca²⁺ release was calculated over the first 100 ms by exponential fitting (Zucchi et al., 1995b). In order to confirm that the quick phase of Ca²⁺ release represented SR Ca²⁺ release, we tested the inhibition of Ca²⁺ release by a ‘non-release' buffer, containing (in mM) HEPES-potassium 20 (pH 6.8), KCl 100, MgCl₂ 10 and 10 μM ruthenium red.

Assay of SR Ca²⁺-ATPase activity and SR Ca²⁺-uptake

ATPase activity was determined in the absence of Ca²⁺ gradient by a coupled enzyme reaction, as described by Feher & Lipford (1985), with minor changes. Briefly, the reaction mixture contained KCl 100 mM, imidazole buffer 20 mM (pH 7.0 at 37°C), phosphoenolpyruvate 1 mM, NADH 0.3 mM, 9 IU ml⁻¹ lactate dehydrogenase, 6 IU ml⁻¹ pyruvate kinase, NaN₃ 5 mM, A23187 2 μM, ATP 1 mM, EGTA 0.2 mM and CaCl₂ 0.25 mM where appropriate. Ca²⁺-dependent ATPase activity was defined as the difference between the activity measured in the presence and in the absence of 0.25 mM CaCl₂.

Oxalate-supported Ca²⁺-uptake was determined in the crude homogenate in the presence of a concentration of ryanodine able to block the Ca²⁺ release channel, as described previously (Zucchi et al., 1994). As discussed elsewhere, oxalate is accumulated into the SR, and therefore such an assay provides a reliable estimate of SR Ca²⁺-uptake, even in crude preparations (Feher & Lipford, 1985). Unless otherwise specified, free Ca²⁺ concentration was 1 μM.

Perfusion experiments

Rats were anaesthetized as described above. After injection of 1000 u sodium heparin in the femoral vein, the heart was quickly excised and perfused according to the working heart technique, as described previously (Zucchi et al., 1994). The preload (height of the atrial chamber) and the afterload (height of the aortic chamber) were set at 20 and 100 cm, respectively. The standard perfusion buffer included (mM): NaCl 118, NaHCO₃ 25, KCl 4.5, KH₂PO₄ 1.2, MgSO₄ 1.2, CaCl₂ 1.5, glucose 11. Perfusions were carried out using 200 ml of recirculating buffer, which was equilibrated with a mixture of O₂ (95%) and CO₂ (5%). Temperature was kept between 36.8 and 37°C, and the pH was 7.4. After 5 min of equilibration, doxorubicin, epirubicin or MEN 10755 were added to the perfusion buffer, and the haemodynamic variables were monitored for 60 min.

Chemical and radionuclides

MEN 10755, MEN 11383 (7-O-[2,6-dideoxy-4-O-(2,3,6-trideoxy - 3-amino -α-L-lyxo-hexopyranosyl)-α- L-lyxo-hexopyranosyl] -4 -demethoxy -13-dihydro -14 - hydroxydaunomycinone hydrochloride, i.e. the 13-dihydro derivative of MEN 10755) and doxorubicinol (13-dihydro-doxorubicin) were provided by Menarini Ricerchi S.p.A (Pomezia, Italy). Doxorubicin (Adriblastina) and epirubicin (Epirubicina) were obtained from commercial preparations. Ryanodine was purchased from Calbiochem (San Diego, CA, U.S.A.). EGTA was obtained from Sigma Chemicals (St. Louis, MO, U.S.A.). All other reagents were of analytical grade. Free Ca²⁺ concentration was calculated according to Fabiato & Fabiato (1979). Free Ca²⁺ was also measured with the antipyrylazo III technique (Scarpa, 1979), and the results of the assay were generally in accordance with the theoretical values. [³H]-ryanodine and ⁴⁵CaCl₂ were obtained from New England Nuclear – DuPont (Milan, Italy).

Statistical analysis

Results are expressed as mean±s.e.mean. Data analysis was performed using GraphPad Prism version 3.00 for Windows (GraphPad Sofware, San Diego, CA, U.S.A.). Binding experiments were analysed using the models mentioned in the description of the specific experiments. Release experiments were analysed by least squares linear regression, after logarithmic transformation (see also Zucchi et al., 1995b). Differences between groups were evaluated as follows. One-way analysis of variance was used a global test for differences between means. If between-groups variance was significantly (P<0.05) higher than within-groups variance, individual groups were compared by Student-Neuman-Keuls test.

Results

[³H]-ryanodine binding experiments

Ryanodine binding is strongly Ca²⁺ -dependent. Ca²⁺ affected both the K_D for ryanodine and the B_max. In our SR preparation, at 1 μM Ca²⁺ ryanodine binding was characterized by K_D=14.3 nM and B_max=450 fmol mg⁻¹ protein, whereas at 30 μM Ca²⁺ the K_D decreased to 2.1 nM and the B_max increased to 4.1 pmol mg⁻¹. The EC₅₀ for Ca²⁺ (determined at 2 nM [³H]-ryanodine; see Figure 1, triangles) averaged 5.1 μM (pEC₅₀=5.15±0.09).

Figure 1.

[³H]-Ryanodine binding was determined at 2 nM [³H]-ryanodine concentration and different free Ca²⁺ concentrations, under control conditions and in the presence of 50 μM doxorubicin or 50 μM MEN 10755. Data points represent mean±s.e.mean of three different experiments. Data was analysed by GraphPad Prism 3.00 software, using a sigmoidal dose-response model. EC₅₀ values averaged 5.1, 1.0 and 2.6 μM in the three groups, respectively (pEC₅₀=5.15±0.09, 5.98±0.03 and 5.58±0.09), and the difference between any pair of groups was statistically significant (P<0.01 by ANOVA and Student-Neuman-Keuls test). Maximum binding averaged 1810±144, 1768±34 and 1487±69 fmol mg⁻¹ protein, respectively. The Hill coefficient was close to 2 in all groups.

The chief effect of doxorubicin was to increase the Ca²⁺-sensitivity of ryanodine binding (Figure 1), by reducing the EC₅₀ to 1.0 μM (pEC₅₀=5.98±0.03). A dose-response curve obtained at fixed concentrations of ryanodine and free Ca²⁺ (2 nM and 1 μM, respectively) is shown in Figure 2. Doxorubicin increased ryanodine binding by up to 15.3±2.5 fold, with EC₅₀=20.6 μM (pEC₅₀=4.69±0.14). Epirubicin, the 3′-epimer of doxorubicin, produced similar effects, namely 9.7±0.6-fold maximum stimulation with EC₅₀=11.1 μM (pEC₅₀=4.95±0.06). Doxorubicinol caused 3.1±0.1-fold stimulation with EC₅₀=3.8 μM (pEC₅₀=5.42±0.08). These effects were due both to increased B_max, and to decreased K_D (data not shown). On the other hand, at 30 μM Ca²⁺ concentration, doxorubicin increased ryanodine binding by less than 50% (Figure 3).

Figure 2.

[³H]-Ryanodine binding was determined at 2 nM [³H]-ryanodine concentration, in the presence of 1 μM free Ca²⁺, and of 0–100 μM doxorubicin, epirubicin, doxorubicinol or MEN 10755. Data points represent mean±s.e.mean of three different experiments and are expressed as ratio to baseline binding. Data was analysed by GraphPad Prism 3.00 software, using a sigmoidal dose-response model. EC₅₀ values averaged 20.6, 11.1, 3.8 and 38.9 μM, respectively (pEC₅₀=4.69±0.14, 4.95±0.06, 5.42±0.08 and 4.41±0.19). Maximum binding stimulation averaged 15.3±2.5, 9.7±0.6, 3.1±0.1 and 1.9±0.3. The difference between the MEN 10755 curve and either the doxorubicin or the epirubicin curve was statistically significant (P<0.01 by ANOVA and Student-Neuman-Keuls test).

Figure 3.

[³H]-Ryanodine binding was determined at 2 nM [³H]-ryanodine concentration in the presence of 30 μM free Ca²⁺, and of 0–100 μM doxorubicin, or MEN 10755. Data points represent mean±s.e.mean of two different experiments and are expressed as ratio to baseline binding.

MEN 10755 produced much lower effects than doxorubicin or epirubicin (Figures 1 and 2). In particular, at 1 μM free Ca²⁺ concentration, MEN increased ryanodine binding by up to 1.9±0.3 fold (P<0.01 vs doxorubicin and epirubicin), with EC₅₀=38.9 μM (pEC₅₀=4.41±0.19). This effect was basically due to increased B_max, whereas the K_D was unchanged (data not shown). MEN 11383, the 13-dihydroderivative of MEN10755, did not affect ryanodine binding (data not shown). At 30 μM Ca²⁺ concentration, no significant increase in ryanodine binding was produced by MEN 10755, even at 100 μM concentration (Figure 3).

⁴⁵Ca release experiments

To confirm the functional implications of the ryanodine binding experiments, we assayed SR Ca²⁺-induced Ca²⁺ release. After in vitro loading, the ⁴⁵Ca²⁺ content of our SR preparation was of the order of 20–25 nmol mg⁻¹ protein. Exposure to the release buffer determined the quick release of 8–12 nmol mg⁻¹, which was completed over 100–120 ms. Thereafter the rate of Ca²⁺ release decreased considerably. The quick component of Ca²⁺ release was inhibited by 10 mM Mg²⁺ plus 10 μM ruthenium red, confirming that it represented Ca²⁺ efflux through the SR release channels (release curves were quite similar to those shown in Zucchi et al., 1995b).

The effects of doxorubicin and MEN 10755 were tested using a release buffer that included 2 μM free Ca²⁺. The extent of ⁴⁵Ca²⁺ loading and of the ⁴⁵Ca²⁺ pool subjected to quick release were similar after treatment with either 50 μM doxorubicin or 50 μM MEN 10755 (releasable Ca²⁺ averaged 10.3±1.5 nmol mg⁻¹ protein, without any significant difference between groups). The rate constant of Ca²⁺ release averaged 22±2 s⁻¹ under control conditions, and it increased by about 4 fold (82±5 s⁻¹, P<0.01 vs control) in the presence of doxorubicin (Figure 4). On the other hand, in the presence of MEN 10755, SR Ca²⁺ release was not significantly different from the baseline (rate constant 24±3 s⁻¹, P<0.01 vs doxorubicin).

Figure 4.

SR Ca²⁺-induced Ca²⁺ release curves were determined with a release buffer containing 2 μM free Ca²⁺, under control conditions and in the presence of 50 μM doxorubicin or 50 μM MEN 10755. The vertical axis shows the logarithm of the ratio of ⁴⁵Ca²⁺ to ⁴⁵Ca²⁺ at time zero, after subtraction of the amount of ⁴⁵Ca²⁺ not involved in quick Ca²⁺ release. Since the release follows exponential kinetics, the slope of the regression lines represents the rate constant of Ca²⁺-induced Ca²⁺ release (for further details see Zucchi et al., 1995b). Data points represent mean±s.e.mean of three different experiments. Data was analysed by GraphPad Prism 3.00 software, using a linear regression model. The rate constant averaged 82±5 s⁻¹ in the presence of doxorubicin and 24±3 s⁻¹ in the presence of MEN 10755, vs a control value of 22±2 s⁻¹. The doxorubicin curve was significantly different from the control and MEN 10755 curves (P<0.01 by ANOVA and Student-Neuman-Keuls test), whereas the difference between the control and MEN 10755 curves did not achieve statistical significance. The amount of Ca²⁺ subjected to quick release averaged 10.3±1.5 nmol mg⁻¹ protein, without any significant difference between groups.

Assay of SR Ca²⁺ -ATPase activity and of oxalate-supported Ca²⁺-uptake

In our microsomal preparation, Ca²⁺-ATPase activity averaged 1.24±0.10 μmol min⁻¹ mg⁻¹ protein. As shown in Table 1, none of the compounds tested (i.e. doxorubicin, doxorubicinol, MEN 10755 and MEN 11383) produced any significant effect at concentrations ⩽100 μM. At higher concentrations, some inhibition occurred with doxorubicinol (42% inhibition at 200 μM concentration, P<0.05). The assay of Ca²⁺-uptake confirmed that no compound was able to produce significant effects at 50 μM concentration, whereas a slight inhibition (14%) was observed with 200 μM doxorubicinol, although it did not reach the threshold of statistical significance.

Table 1.

SR Ca²⁺-ATPase and Ca²⁺-uptake

Perfusion experiments

The results of perfusion experiments are summarized in Figure 5. Baseline values of the chief haemodynamic variables were similar in all experimental groups, and they averaged as follows: aortic flow 48.5±1.9 ml min⁻¹, coronary flow 16.9±0.9 ml min⁻¹, cardiac output 65.4±2.7 ml min⁻¹, peak systolic aortic pressure 124±3 mmHg, heart rate 265±13 beats min⁻¹. In the control group, that was subjected to 60 min of aerobic perfusion, all haemodynamic variables were substantially stable (changes were <10% in all cases). The addition of 30 μM doxorubicin to the perfusion buffer produced a progressive decrease in aortic flow, coronary flow, cardiac output and peak aortic pressure. After 50–60 min, half of the hearts failed (i.e. they were no longer able to sustain a significant aortic flow). At the end of the perfusion, the mean values of all haemodynamic variables were <30% of the baseline. Epirubicin (30 μM) produced similar or even greater contractile impairment, since all the hearts of this group failed before 50 min. On the other hand, perfusion with 30 μM MEN 10755 produced only minor effects. After 60 min of perfusion, aortic flow and cardiac output averaged 85±13 and 93±8% of the baseline, respectively, whereas coronary flow, peak aortic pressure and heart rate were virtually unchanged. Statistical analysis showed that the effects of MEN 10755 were significantly different from those of doxorubicin or epirubicin (P<0.01 in both cases), while the difference between the control group and the MEN 10755 group was not significant.

Figure 5.

Time course of the chief haemodynamic variables in hearts perfused under control conditions or in the presence of 30 μM doxorubicin, 30 μM epirubicin, 30 μM MEN 10755. (a) aortic flow; (b) cardiac output; (c) peak aortic pressure; (d) heart rate; (e) coronary flow. Data points are expressed as percentage of the baseline values, and represent mean±s.e.mean of four hearts in each group. Actual baseline values are given in the text. ^*P<0.01 vs the control group and the MEN 10755 group, by ANOVA and Student-Neuman-Keuls test. Differences between the control group and the MEN 10755 group were not statistically significant for any variable, at any time.

Discussion

MEN 10755 is a disaccharide analogue of doxorubicin, which appears to be more effective than doxorubicin in the inhibition of topoisomerase II and in the treatment of human tumour xenografts (Arcamone et al., 1997; Pratesi et al., 1998). Since the clinical use of anthraquinone compounds is often limited by cardiac toxicity (Buzdar et al., 1985; Jensen, 1986; Hagane et al., 1998; Doroshow, 1991; Rhoden et al., 1993; Wang & Korth, 1995; Nysom et al., 1998), it seemed interesting to evaluate the cardiac effects of MEN 10755, which were compared with those of doxorubicin and epirubicin.

The cardiac toxicity of anthraquinones does not seem to be related to topoisomerase II inhibition. There is strong evidence that a major role is played by impaired intracellular Ca²⁺ homeostasis (Kusuoka et al., 1991; Jiang et al., 1994; Temma et al., 1994; Halili-Rutman et al., 1997; Maeda et al., 1998). In particular, acute cardiac injury is likely to be due to stimulation of Ca²⁺ release through the SR Ca²⁺ channel. Doxorubicin acts by sensitizing the channel to the stimulatory action of extravescicular Ca²⁺. As it is the case for many activators of the SR Ca²⁺ channel, doxorubicin increases the binding of ryanodine, a selective ligand that can only interact with the open state of the channel (Zorzato et al., 1985; Abramson et al., 1988; Kim et al., 1989; Ondrias et al., 1990; Pessah et al., 1990; Tian et al., 1991; Boucek et al., 1993; Zucchi & Ronca-Testoni, 1997). These effects were confirmed in the present investigation, since at micromolar Ca²⁺ concentration doxorubicin determined a remarkable increase in ryanodine binding and in the rate of Ca²⁺-induced Ca²⁺ release. These actions were shared by epirubicin, the 3′-epimer of doxorubicin.

MEN 107551 produced only minor effects on the ryanodine receptor. At 30 μM free Ca²⁺ concentration, ryanodine binding was unaffected by up to 100 μM MEN 10755. At 1 μM free Ca²⁺, the binding of 2 nM ryanodine was only slightly increased. The maximum stimulation was on the order of 1.9 fold (vs 15.3 fold in the case of doxorubicin). The results of release experiments were in accordance with these findings, since 50 μM MEN 10755 did not produce any significant change in the rate of SR Ca²⁺ release, at 2 μM free Ca²⁺ concentration.

The interaction between anthaquinones and the ryanodine receptor has not been characterized at molecular level. Anthraquinones are thought to bind to a site that overlaps at least in part with the caffeine binding site (Abramson et al., 1988; Pessah et al., 1990), and to modify hyperreactive cysteine residues (Feng et al., 1999). Our results suggest that the saccharidic moiety of the molecule and/or the 3′-methoxy group play a major role in the binding reaction, or in the modulation of channel function.

Some investigators have suggested that doxorubicin toxicity may be due to inhibition of the SR Ca²⁺-ATPase. In particular, ATPase inhibition by doxorubicinol (13-dihydrodoxorubicin), a catabolite of doxorubicin, has been reported (Boucek et al., 1987; Olson et al., 1988). In our model, none of the compounds tested (doxorubicin, doxorubicinol, MEN 10755 and its 13-dihydro-derivative) significantly affected Ca²⁺-ATPase activity, or Ca²⁺ uptake, at concentrations ⩽100 μM. Some inhibition occurred only with higher concentrations of doxorubicinol. Therefore, it seems unlikely that interference with SR Ca²⁺ uptake may be relevant in vivo.

Since activation of SR Ca²⁺ release is believed to be the chief cause of acute anthraquinone toxicity, MEN 10755 should be much less toxic than doxorubicin or epirubicin. To test this prediction, we evaluated the acute effects of these compounds in an isolated heart model. The difference between 30 μM MEN 10755 and equimolar doxorubicin or epirubicin was dramatic. After 60 min of perfusion, serious contractile impairment occurred in every heart treated with doxorubicin or epirubicin, whereas in the MEN 10755 group all haemodynamic variables were close to the normal range. These findings are in good agreement with the results reported by Parlani et al. (1998) in rat right ventricular muscle, and induce to believe that in the clinical setting the acute administration of MEN 10755 should be much less cardiotoxic than observed with doxorubicin or epirubicin.

The pathogenesis of chronic doxorubicin toxicity is poorly understood. The same mechanisms that are responsible for the acute effects might be involved. In particular, low doses of doxorubicin might produce marginal cellular injury (i.e. necrosis limited to a few cardiomyocytes), and doxorubicin cardiomyopathy might represent the cumulative result of these lesions. However, additional mechanisms are likely to play a role. After prolonged exposure to doxorubicin, ryanodine receptor activation is followed by a slowly-developing irreversible inactivation (reviewed in Zucchi & Ronca-Testoni, 1997). Reduced density of SR Ca²⁺ release channels has actually been observed in models of chronic doxorubicin toxicity (Pessah et al., 1992; Dodd et al., 1993), and the expression of several genes coding for SR proteins–ryanodine receptor, calsequestrin, SR Ca²⁺-ATPase and phospholamban–is decreased (Arai et al., 1998). Such long-term changes in SR proteins might reduce the availability of Ca²⁺ for contractile protein activation, contributing to inotropic failure. Consistently, measurement of intracellular Ca²⁺ after prolonged exposure to doxorubicin has shown reduced amplitude and prolonged duration of Ca²⁺ transients (Jiang et al., 1994; Temma et al., 1994; Maeda et al., 1998), while the Ca²⁺-accumulating capability of the SR was decreased (Halili-Rutman et al., 1997).

On the basis of the remarkable difference which we have observed between the SR effects of doxorubicin and MEN 10755, one could expect lower toxicity of the disaccharide analogue even after chronic administration. Recent experimental results appear to support this hypothesis (Cirillo et al., 2000), but specific studies on SR changes after prolonged exposure to MEN 10755 are still required for a thorough evaluation of this important issue.

Abbreviations

MEN 10755

7-O-[2,6-dideoxy-4-O-(2,3,6-trideoxy-3-amino-α-L-lyxo-hexopyranosyl)-α-L-lyxo-hexopyranosyl]-4-demethoxy-14-hydroxydaunomycinone hydrochloride

MEN 11383

7-O-[2,6-dideoxy-4-O-(2,3,6-trideoxy-3-amino-α-L-lyxo-hexopyranosyl)-α-L-lyxo-hexopyranosyl]-4-demethoxy-13-dihydro-14-hydroxydaunomycinone hydrochloride

SR

sarcoplasmic reticulum