Degradation of Anthracycline Antitumor Compounds Catalysed by Metal Ions

The influence of some metal ions on the degradation of anthracyclines was examined. One of the degradation products is the 7,8-dehydro-9,10-desacetyldoxorubicinone, D* (¥), usually formed by hydrolysis at slightly basic pH. D* is a lipophilic compound with no cytostatic properties. Its formation could be responsible for the lack of antitumor activity of the parent compound. The coordination of metal ions to anthracycline derivatives is required to have degradation products. Cations such as Na+, K+, or Ca2+ do not induce the D* formation however metals which can form stable complexes with doxorubicin afford D*. Iron(III) and copper(II) form appreciable amount of D* at slightly acidic pH. Terbium(III) forms D* but its complex is stable only at slightly basic pH. Palladium(II) which does not form D*. The influence of the coordination mode of metal ions to anthracycline on the D* formation is discussed.


Introduction
Degradation of antitumor compounds is a common problem for pharmacologists. 1 The efficacy of an antitumor drug is related to the intracellular drug concentration as well as to the rate of cell uptake. 2 Anthracycline antibiotics, Doxorubicin (I) and Daunorubicin (l')(Chart 1) are benzanthraquinone drugs, which are useful in the treatement of several type of human malignancies. 3  Complexation of these biological active molecules with metal ions has been examined in order i) to reduce the toxic effects such as dose-dependent cardiotoxicity; 5 ii) to improve the antitumor properties: 6 iii) to reproduce some in vivo mechanisms . 7 In order to determine the relationships between intracellular drug concentration and the cytotoxic activity (as well as the degree of resistance) we were interested in the anthracycline degradation in cell culture.
One of the degradation products of anthracycline, namely 7,8-dehydro-9,10-desacetyldoxorubicinone, D* (ll)(Charl; 2), has been idemified 8'9 however its biological effects have been neglected by pharmacologists. In a previous report we have shown that this derivative, which is detected in cells following incubation with doxorubicin, is formed in the culture medium. We also detected the presence of D* when studying the interactions of cells with the iron complexes of doxorubicin and pirarubicin; in this case the degradation product is not formed in the medium but it is already present in the metal complex solution. 1 1 These outcomes prompted us to further investigate whether other metal ions can form D* giving us the opportunity to discuss about the chemical requirements necessary to have D* formation.

Results
The metal complexes of anthracycline were prepared following the published procedures 12-15 and dissociated by adding hydrochloric acid (0.2 M). The dissociation step usually required less than one hour except for the Pd(ll) complexes. We verified that the strongly acidic pH (pH<l) was not responsible of the degradation of anthracyclines. Control experiments were done with the free anthracycline and D* was not formed in the experimental conditions employed without metal ion. Usually the iron complexes are prepared starting from iron(ll), in the form of Mohr's salt as source of iron(Ill), following air-oxidation to iron(Ill). This method is generally considered the best way to prepare these complexes and to avoid formation of iron hydroxides. 8 The metal ion can be added either to an anthracycline buffer solution at pH 7 (method i) or to an unbuffered solution and the Volume 1, Nos. [2][3]1994 Degradation of Anthracycline Antitumour Compounds Catalysed by Metal Ions subsequent addition of the stoichiometric amount of NaOH necessary to deprotonate once the anthracycline (ii). Both reactions are very fast and depending on the procedures used, two different CD spectra were obtained. In buffer solution (method i) and also in the presence of an excess of iron, the CD spectrum is prevalently characterized by a strong positive band centred at 630 nm ( Fig.   1)" this band is typical of the iron complexes of daunorubicin derivatives 2,19 and characterizes what will be hereafter labelled a D-type spectrum. 24 hours later, after dissociation of the complex, we detected the presence of D* in an amount ranging from 0 to 10% of the initial anthracycline concentration.
If the metal complexes are prepared in water with the subsequent addition of NaOH, 24 hours later the amount of D* raised to 25-30%: the CD spectrum did not evolve with time and still exhibited two positive bands at 500 and 605 nm, characteristics of an A-type spectrum ( This complex is formed in water at pH around 5-6; at these values the hydrolytic pathway to get D* is discouragead. 8 In fact in the same conditions, free doxorubicin did not yield D*.

Discussion
The instability of anthracycline derivatives in aqueous solution and in cell culture media has been recognised by several investigators.
Among them Beijnen eta/. 8,9,26 have identified the anthracycline degradation products formed in acidic media (pH<4) or slightly basic pH media (pH~8). They showed that degradation of doxorubicin at pH higher than 7 yielded mainly a pink coloured aglycone in which full aromatization of the ring A occurred. This compound, identified as D*, can be also formed at strongly acidic pH (pH<1)27 and by photodegradation. 28 Structural analogies of the doxorubicin with corticosteroids, i.e. hydrocortisone, suggested a possible mechanism of the D* formation at basic pH. 8 Corticosteroids possess a C17 (-ketol side chain 29 analogue to the C9 (-ketol substituent in doxorubicin which could constitute an extra protonation site. Reversible enolization and ionization of this function should be possible at basic pH: the determination of the corresponding pKa would be hampered by overlap with the deprotonation of the phenolic group. In analogy with the degradation of corticosteroids, the enolization step is assumed to be the first step of the degradation reaction of doxorubicin. The enolate anion arising from keto/enol tautomerization and deprotonation may be involved in a tautomeric equilibrium with its 13-o1-14-aidehyde derivative. A concertated mechanism entraining the cleavage of the C9 side chain and the aminosugar moiety yields D*. The high stability of the resulting degradation product is the driving force of the full aromatization of the A ring after cleavage of the C9 side chain. D*, a lipophilic compound, enters in cells very rapidly but unfortunately it has no cytostatic properties:l I its formation could explain the reduced antiproliferative activity observed in the case of degradated anthracycline. 1 0 An influence of metal ions on the D* degradation was already proposed 26 but not verified. In the case of the Tb(lll)complexes, the slightly basic pH at which this complex is formed could be considered as responsible of the doxorubicin degradation. It has to be observed that at the same pH value, free doxorubicin is more degrated (50% in 1 day) than if complexed to terbium(Ill) ions. It means that in this case the metal ion seems to have a protecting role against the D* formation.
However from our results the formation of D* seems to be related to the different coordination mode of the metal ions. In fact coordination at the C5-C6 site, which has been established for Cu(ll) 21 and Tb(lll) 14 seems to prevent D* formation more than the coordination at the Cll-C12 site. The anomalous case of terbium(Ill) has been related to the high value of the reactionnel pH.
The C11-C12 position has probably a positive effect on the enol formation, first step of the degradation pathway.
In addition under slightly basic conditions the rate of the glycosidic bond cleavage is strongly dependent on the electronic structure of the aglycone part 8 while at acidic pH the structural modifications in the sugar moiety are more important. 26 Palladium(ll) is coordinated to doxorubicin also by the amino group of the daunosamine and, in this case D* has been never observed.
This important difference in the coordination mode of metal ion seems to play a role in D* formation.
This fact could be related to the cleavage of the glycosidic bond which is a crucial step in the concertated mechanism yielding the D* formation.
From a recent paper 31  To justify the anisotropy of this complex two "speculative" models of polymer were proposed: in the first one [Fe(Dnr)3] octahedra are attached to each other by Dnr-Dnr stacking forming a three dimensional polymeric structure. In the second case a linear polymer would be composed of planes of two Dnr molecules binding Fe(lll) ions with the oxygens at Cll and C12. The axial positions would be occupied by OH groups bridging two planes. Planes of Dnr dimers would be randomly distributed in the linear polymer with in both cases the amino group of the daunorubicin involved in the stacking of the anthracyclines.
Making the hypothesis that the structure of the [Fe(Dox)3] complex exhibiting a D-type CD spectrum is similar to that of the [Fe(Dnr)3] we can suggest that the involvment of the amino group in the Dox stacking prevent D* formatio as it was observed in the case of the palladium complex.
Obviously these hypothesis require to be verified by the analysis of other complexes of anthracycline, which are actually under study.