by scientists in the Swiss Federal Institute of Technology in. Zurich and Ciba in ... term DFO therapy in thalassaemia, and, finally, review the development of new ...
British Journal of Haematology, 1998, 101, 399–406
Review IRON CHELATORS FOR THALASSAEMIA The iron-chelating compound desferrioxamine (DFO) was discovered accidentally, as a byproduct of antibiotic research by scientists in the Swiss Federal Institute of Technology in Zurich and Ciba in Basel. Its development into a useful clinical drug was not by design, but thanks to natural curiosity and ingenuity, as described by Keberle (1992). Although DFO has been available for treating transfusional iron overload from the early 1960s, the era of modern and effective iron-chelating therapy started only 20 years ago with the introduction of subcutaneous DFO infusions by portable pumps. Today, long-term DFO therapy is an integral part of the management of thalassaemia and other transfusion-dependent anaemias, with a major impact on wellbeing and survival. In this Review we will discuss the abnormalities in iron metabolism associated with thalassaemia major and the chelatable iron pools representing the targets of iron-chelating therapy, examine the results of longterm DFO therapy in thalassaemia, and, finally, review the development of new, orally effective, iron chelators which may be considered for present or future use. Pathophysiology of thalassaemic iron overload Iron balance in normal subjects is characterized by an efficient recycling of wastage iron from senescent erythrocytes. Only a fraction of the daily iron requirement is obtained by intestinal absorption (Cook et al, 1970). The primary abnormality in thalassaemia major is a wasteful ineffective erythropoiesis resulting in a 10–15-fold expansion of the erythroid bone marrow (Hershko & Rachmilewitz, 1979) and a drastic increase in haemoglobin catabolism. Iron accumulation is the consequence of blood transfusions as well as of increased iron absorption caused by erythropoietic activity. The role of increased iron absorption is illustrated by the severe iron overload encountered in patients with thalassaemia intermedia who have never received blood transfusions. The combination of iron overload and increased outpouring of catabolic iron from the reticuloendothelial system overwhelms the iron-carrying capacity of transferrin, resulting in the emergence of toxic non-transferrin bound plasma iron (NTPI). Although our original description of a chelatable, low molecular weight, plasma iron fraction in thalassaemic patients with severe iron overload (Hershko et al, 1978) was greeted with initial scepticism, its existence has been confirmed by subsequent studies using a variety of methods. NTPI promotes the formation of free hydroxyl radicals and accelerates the peroxidation of membrane lipids in vitro Correspondence: Dr C. Hershko, Department of Medicine, Shaare Zedek Medical Centre, P.O. Box 3235, Jerusalem 91031, Israel. q 1998 Blackwell Science Ltd
(Gutteridge et al, 1985). Long-term treatment with DFO or deferiprone (L1) in thalassaemic patients results in a marked decrease in their NTPI measurements (Al-Refaie et al, 1992). More recently, Porter et al (1996) have shown that plasma NTPI is removed by intravenous DFO therapy in a biphasic manner and that upon cessation of DFO infusion it reappears rapidly, lending support to the continuous, rather than intermittent, use of DFO in high-risk patients. The rate of low molecular weight iron uptake by cultured rat heart cells is over 300-times that of transferrin iron (Link et al, 1985). Moreover, transferrin-iron uptake is inhibited at high tissue iron concentrations due to the down-regulation of transferrin receptor production, whereas non-transferrin iron uptake is actually increased at high tissue iron concentrations (Randell et al, 1994). Such uptake results in increased myocardial lipid peroxidation and abnormal contractility, and these effects are reversed by in vitro treatment with DFO (Link et al, 1989). Recognition of NTPI as a potentially toxic component of plasma iron in thalassaemic siderosis has important practical implications for designing better strategies for the effective administration of DFO and other iron chelating drugs. Without iron-chelating therapy, the accumulation of iron progresses relentlessly, and when about 20 g of iron have been retained in the body, significant clinical manifestations of iron toxicity may be anticipated (Gabutti & BorgnaPignatti 1994). The most important complications of transfusional siderosis are cardiac, hepatic and endocrine disease. Pathologic findings in the heart include dilated, thickened ventricular walls with particularly heavy iron deposits in the ventricles, epicardium and papillary muscles. These cellular deposits induce increased membrane lipid peroxidation in the sarcolemma resulting in impaired Na,K,ATPase activity (Link et al, 1994), increased lysosomal fragility (Link et al, 1993) and, in particular, impaired mitochondrial inner-membrane respiratory chain activity (Link et al, 1996). It is possible to demonstrate early myocardial dysfunction in asymptomatic patients using a multi-gated imaging (MUGA) scan (Aldouri et al, 1990) or dobutamine stress echocardiography (Mariotti et al, 1993). Progressive cardiac siderosis leads to the development of symptomatic heart failure and life-threatening arrhythmias. Myocardial siderosis is the single most important cause of mortality in inadequately-treated thalassaemics. Hepatic fibrosis is a common complication of thalassaemia and, similar to cardiac problems, its incidence is age-related. It has been proposed that patients with liver iron concentrations < 15 mg/g dry weight may have a prolonged survival free of the clinical complications of iron overload (Brittenham et al, 1994). Iron overload per se is responsible for the
development of cirrhosis in many cases. However, the coexistence of chronic hepatitis B or C with an incidence ranging from 8·7% to 70% of thalassaemic patients in various geographic areas (Gabutti & Borgna-Pignatti, 1994) underlines the complexity of this problem. Endocrine problems caused by direct accumulation of iron in endocrine glands or indirectly through the hypothalamic–pituitary axis are common. Stunted growth, delayed puberty, hypothyroidism, hypoparathyroidism and diabetes mellitus are all well-established complications of transfusional siderosis (De Sanctis et al, 1992). Because diabetes and hypothyroidism appear when most endocrine cells are destroyed and replaced by fibrosis, these complications are rarely reversible. Chelatable iron pools The intracellular labile iron pool. Current models of iron acquisition, sequestration and storage by mammalian cells are based on a regulated adjustment of membrane transferrin receptor and cytosolic ferritin levels. Iron in transit between these two iron-binding proteins is believed to exist in a weakly bound low-molecular-weight complex which is also available for interaction with iron chelating drugs (Rothman et al, 1992). This chelatable labile iron pool (LIP) is assumed to be sensed by a cytosolic iron-responsive protein (IRP) which coordinately represses ferritin mRNA translation and increases transferrin receptor mRNA stability (Klausner et al, 1993). Efficient regulatory mechanisms prevent fluctuations in the size of the labile iron pool under conditions of moderate iron deprivation and iron loading. However, massive iron loading results in uncontrollable expansion of the chelatable pool, which fails to be matched by the sequestrating capacity of cellular ferritin (Breuer et al, 1997). This expanded LIP is an obvious target of intracellular iron chelation by drugs that are able to cross the barrier of the cytoplasmic membrane. Role of iron stores in reticuloendothelial and parenchymal cells. Although excess iron may be deposited in almost all tissues, most of it is found in association with two cell types: reticuloendothelial (RE) cells in the spleen, liver and bone marrow, and parenchymal tissues represented by hepatocytes, endocrine cells and the myocardium. In contrast to RE cells in which iron accumulation is relatively harmless, parenchymal siderosis may result in significant organ damage. The source of iron and the proportion retained in ferritin stores or recycled into the circulation from the two cell types is quite different. RE cells have a limited ability to assimilate transferrin iron and they derive iron from the catabolism of haemoglobin in non-viable erythrocytes (Hershko & Weatherall, 1988). Most of this catabolic iron is recycled to plasma transferrin or NTPI within a few hours. In contrast, hepatic parenchymal cells maintain a dynamic equilibrium with plasma transferrin, with iron uptake predominating when transferrin saturation is high, and release when serum iron and transferrin saturation are low. Unlike RE cells, the turnover of parenchymal iron stores is extremely low. In general, iron overload associated with increased intestinal absorption such as hereditary haemochromatosis results in predominant parenchymal siderosis,
whereas in iron overload caused by multiple blood transfusions the primary site of siderosis is the RE cells. Considerable redistribution of iron may take place subsequently. Experimental and clinical observations indicate that the urinary excretion of chelated iron is derived mainly from RE cells. Studies in hypertransfused rats have shown that, in contrast to hepatocellular radioiron excretion which is confined entirely to the bile, most of the radioiron excretion derived from RE cells is recovered in the urine. Moreover, when DTPA or IRC11, water-soluble synthetic chelators which do not enter cells easily, are employed in the same experimental model, there is no enhancement at all of hepatocellular iron excretion, but the enhancement of urinary RE radioiron excretion is similar or higher than that observed previously with DF (Rivkin et al, 1997). Hence, DF obtains iron for chelation by one of two alternative mechanisms: (a) in situ interaction with hepatocellular iron and subsequent biliary excretion, and (b) chelation of iron derived from red blood cell catabolism in the RE system directly or following its release into the plasma in the form of NTPI with subsequent urinary excretion. Results of long-term chelating therapy The introduction of desferrioxamine for the treatment of transfusional iron overload had a major impact on the survival and well-being of thalassaemic patients. Although this has never been proven by prospective, randomized clinical studies, and such randomized studies are no longer justified ethically, the beneficial effects of long-term desferrioxamine treatment are clearly demonstrated by comparison of treatment outcome with historic controls. Experience with long-term DFO therapy in thalassaemic patients has been the subject of a number of excellent recent reviews (Gabutti & Borgna-Pignatti, 1994; Gabutti & Piga, 1996; Olivieri & Brittenham, 1997). Method of treatment. The effect of treatment, measured by urinary iron excretion, is directly proportional to the severity of iron overload. Hence, treatment in subjects without iron overload will result in limited iron excretion. However, treatment should not be introduced too late if the objective is the prevention of iron toxicity. It should be started when serum ferritin levels reach about 1000 mg/l which usually occurs after the first 10 or 20 transfusions (Gabutti & Piga, 1996). DFO is infused via a thin s.c. needle in the arm or abdomen at night, connected to a portable pump over 8– 12 h, five to seven times per week, at a daily dose of 20– 60 mg/kg. A urinary iron excretion of 0·5 mg/kg/d is usually sufficient to ensure negative iron balance. Although prolonged subcutaneous DFO infusion is universally recognized as the optimal method of treatment, twice daily s.c. injections may yield similar amounts of urinary iron (Borgna-Pignatti & Cohen, 1997). A new delivery system for continuous DF infusion has been introduced by Baxter enabling continuous 48 h s.c. or continuous 24 h i.v. delivery for 7 d each week (Araujo et al, 1996). This technology enables effective removal of toxic free iron (NTPI) from the plasma, a significant decrease in serum ferritin within 4 weeks, and improves patient compliance compared to conventional s.c. DFO pumps.
q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 399–406
Review Recently, a depot preparation of DFO has been developed by Novartis which, by delivering a smaller dose over a longer period of time, makes it more efficient and reduces the proportion of non-chelated wasted DFO, which can be up to 90% of the drug when delivered by current methods (Alberti et al, 1997). Preliminary studies have shown that the duration of action of the depot formulation is >30 h, and cumulative iron excretion is >3 times that of standard s.c. bolus DFO injections. If local tolerability of the new depot preparation in current clinical studies proves to be acceptable, this new technology may permit once-daily or alternate-day s.c. injections, obviating the need for portable pumps and improving patient compliance. Response to treatment may be assessed by serum ferritin measurements, liver biopsies, computed tomography, or magnetic susceptibility (SQUID) (Brittenham et al, 1982). Serum ferritins are disproportionately low in patients with coexistent ascorbate deficiency (Chapman et al, 1982) and high in active liver disease or inflammation. Nevertheless, serum ferritin is the most accessible and inexpensive tool for the long-term monitoring of chelating efficiency and protection from cardiac complications may be achieved when ferritin levels are kept 2500 mg/l, the threshold identified in the above study for the risk of developing cardiac disease. This failure to achieve a steady decrease in storage iron is explained by the difference in efficacy between the two drugs on a weight per weight basis. As shown by a recent metabolic balance study comparing combined urinary and faecal iron excretion in thalassaemic patients receiving either 60 mg/kg DFO or 75 mg/kg p.o. L1, mean iron excretion in L1 patients was only 65% of those on DFO (Grady et al, 1996). However, in some patients L1 was as effective or better than DFO. The results of L1 therapy in non-thalassaemic patients such as in myelodysplasia (MDS) are quite similar to those with thalassaemia: negative iron balance was achieved in 56% of patients, and the main causes of discontinuation of treatment were nausea and arthralgia (Kersten et al, 1996). An unexpected bonus of iron chelating therapy with L1 or DFO in MDS is a significant decrease in transfusion requirement and increased erythroid activity documented by serum transferrin receptor levels, attributed to improved endogenous erythropoietin production (Jensen et al, 1996).
q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 399–406
Review Collectively, these data imply that oral L1 treatment alone will not ensure sufficient protection in all patients and that careful monitoring is required to identify patients in whom additional conventional chelating treatment with DFO is indicated. Concerns related to the accelerated development of hepatic fibrosis have been expressed based on observations made on patients on long-term L1 therapy at the Toronto Hospital for Sick Children (Olivieri, N.F., unpublished observations). This important issue will probably determine the future of L1 in clinical medicine and needs urgent clarification by reviewing the hepatic status of all other patients on long-term L1 treatment. The results of hepatic histology and liver iron measurements in the Swiss group of 11 thalassaemic patients who have been on continuous L1 therapy since 1989 are eagerly awaited (Tondury et al, 1998). An important cause of limited efficiency of deferiprone in humans is its rapid conjugation with glucuronide inhibiting the iron-binding function of the 3-hydroxyl group. Hence, present ligand design is aimed at the development of compounds resisting inactivation by extensive conjugation with glucuronide or sulphate (Hider et al, 1996). The efficacy/toxicity ratio of hydroxypyridone derivatives may be improved by introducing bulky substitutes on the 2 position of the pyridone ring. Such compounds show a limited interaction with non-haem iron enzymes such as lipoxygenase and ribonucleotide reductase, are less active in inducing apoptosis, but retain their ability to mobilize intracellular iron effectively (Porter et al, 1997). The polyanionic amines (HBED) The search for improved, orally effective, iron-chelating compounds has led to the rediscovery of a very powerful polyanionic amine synthesized over 30 years ago by L’Eplattenier et al (1967): N,N0 -bis (2-hydroxybenzoyl)-ethylenediamine-N,N’-diacetic acid (HBED). It forms a hexadentate ligand with ferric iron by its secondary or tertiary nitrogens and hydroxyl and carboxyl groups. The affinity constant for iron(III) of HBED is 39·6, and its affinity for other metals is relatively low. Conversion of the carboxylic groups of HBED to methyl esters results in a marked improvement in its intestinal absorption and a further increase in iron excretion. HBED and dimethyl-HBED are remarkably non-toxic and their LD50 in rats is >800 mg/kg (Hershko et al, 1984). Our studies in hypertransfused rats have shown that at the dose range of 25–50 mg/kg HBED and dimethyl HBED were 12– 15 times more effective than DFO. Iron balance studies performed in a small number of thalassaemic patients treated with HBED (Grady et al, 1994) have shown increased urine and stool iron excretion following oral treatment. Daily total iron excretion with 40 mg/kg/d HBED was 6–11 mg, but no further increase in excretion was achieved at a dose of 80 mg/kg/d. These limited amounts of iron excretion represented 28–48% of the excretion needed for achieving a negative iron balance. Because the prodrug dimethyl-HBED showed improved absorption and bioavailability in animal studies, this compound is now regarded as a promising candidate for clinical evaluation. Likewise, a new prodrug formula of HBED has been developed by Novartis and, in
company with two other orally effective compounds developed by the same group, is presently undergoing intensive evaluation (Schnebli et al, 1997). The substituted polyaza compounds (IRCO11) The substituted cyclic polyaza compound IRC011 was synthesized by Israel Resources Corporation Ltd, Haifa Israel (Winchell et al, 1996). IRCO11 has a number of promising features: (a) it is a hexadentate chelator binding Fe(III) at a 1:1 ratio and therefore without the risk of possible toxic intermediate complexes, in contrast with L1 which is a bidentate chelator requiring three molecules of the drug for each Fe(III) ion; (b) IRC011 is water soluble and its stability constant with Fe(III) is >1000 times that of DFO; (c) its acute LD50 in mice is >4·0 mM/kg compared to 0·44 mM/kg for DF; and (d) unlike DF, polyaza compounds with the structure of IRC011 do not contain any readily hydrolysable covalent bonds, and are anticipated to resist in vivo biotransformation. In contrast to DFO, which promotes the excretion of both hepatocellular and RE iron, the major source of iron mobilized by IRC011 is iron derived from the catabolism of red cell haemoglobin in the RE system (Rivkin et al, 1997). Because in thalassaemia the contribution of red cell (RBC) catabolism to the influx of iron to the plasma is incomparably higher than that of all other sources combined, it is reasonable to assume that RBC catabolism is the predominant source of the toxic and readily chelatable nontransferrin bound plasma iron (NTBI) compartment. Thus IRC011 with its superior ability to interact with catabolic RBC iron may not only improve the rate of urinary iron excretion in thalassaemia, but possibly also offer better protection of vital tissues from peroxidative damage by preventing the formation of, or by eliminating, NTBI. Since IRCO11 is a water-soluble compound, its membrane permeability is limited but it may be regarded as a useful parent compound to a wide spectrum of polyaza analogues. Substitution of its synthetic arms by more lipophilic ligands may result in improved interaction with hepatocellular iron stores, and possibly also better oral activity. Such studies are presently under way. Current issues in chelator development Undoubtedly DFO remains the drug of choice for the management of transfusional siderosis in thalassaemic patients. However, its high cost and the inconvenience of its parenteral administration by portable pumps are major limitations underlining the need for developing alternative orally effective new iron chelating drugs. In the search for new and improved chelators, it is useful to remember some of the basic principles determining the safety and efficacy of iron chelators as defined in an important recent review by Hider et al (1996). (a) Stability of the chelator–iron complex. This is determined not only by the affinity of the drug to iron(III), but also by the nature of their interaction. Hexadentate chelators bind iron at a 1:1 ratio forming a neutral stable complex which prevents iron from participation in harmful reactions producing hydroxyl radicals. In contrast, bidentate chelators
q 1998 Blackwell Science Ltd, British Journal of Haematology 101: 399–406
such as L1 require three molecules to one ferric iron to form a stable neutral complex (Porter et al, 1989a). Hence, at suboptimal concentrations of L1, when either tissue iron concentrations are very high or drug concentrations too low, incomplete 1:1 or 2:1 complexes may be formed which may result in enhanced mobilization of potentially toxic iron. Thus, hexadentate chelators would, on theoretical grounds, be generally preferable to bidentate or tridantate chelators. (b) Partition coefficient, i.e. the relative solubility in water and lipids, determines the ability to cross lipid membranes. Chelators with partition coefficients of 0·05–1·0 will have efficient gastrointestinal absorption. Increasing lipophilicity will result in efficient drug clearance from the portal circulation by the liver. Unfortunately, increasing lipohilicity also improves the penetration of the blood–brain and placental barriers and increases drug toxicity (Porter et al, 1990). (c) Molecular weight is a critical feature in designing new chelators with improved oral efficacy. New chelators with excellent in vitro interaction with chelatable iron in cell cultures, or following their parenteral administration, may be ineffective if their size interferes with their intestinal absorption. To achieve >70% absorption, the molecular weight should be