Pharmacological activities of copper(II) complexes are a direct function of the nature of their ligands associated with the metal ion in vivo. Some of these, defined as OH-inactivating ligands (G. Berthon, Agents Actions 39 (1993) 210–217), may act as specific “lures” for hydroxyl radicals at inflammatory sites and behave as pseudo-catalase-like agents. This property has been advanced for anthranilic acid (H. Miche, V. Brumas, G. Berthon, J. Inorg. Biochem. 68 (1997) 27–38). With a view to improve the chemical features required to render such inactive substances effective anti-inflammatory drugs through their association with copper(II), an in vitro investigation into copper(II) interactions with the anionic form of an anthranilic acid derivative, namely 3-methoxyanthranilate (Man), has been performed under experimental conditions pertaining in vivo. Copper(II)–Man complex equilibria have been determined using glass electrode potentiometry, then checked by UV–vis and mass spectrometries. Given the prime role of histidine as a copper(II) ligand in blood plasma, copper(II)–histidine–Man ternary equilibria have also been studied. Subsequent computer simulations of the distribution of copper(II) in the extracellular fluid revealed that Man can specifically mobilize Cu(II) ions under inflammatory conditions without affecting their distribution under normal physiological conditions. Thiobarbituric acid reactive substances (TBARS) tests conducted with respect to standardized copper-mediated Fenton-type reactions (P. Maestre, L. Lambs, J.P. Thouvenot, G. Berthon, Free Rad. Res. 20 (1994) 205–218) have shown that, like anthranilic acid, Man can effectively both increase the Fenton-like reactivity of copper and decrease the amount of TBARS detected in solution, i.e., act as a potential OH-inactivating ligand.

Metals are found associated with beta-pleated sheets of Abeta42 in vivo and may be involved in their formation. Metal chelation has been proposed as a therapy for Alzheimer's disease on the basis that it may safely dissolve precipitated Abeta peptides. We have followed fibrillisation of Abeta42 in the presence of an additional metal ion (Al(III), Fe(III), Zn(II), Cu(II)) over a period of 32 weeks and we have investigated the dissolution of these aged peptide aggregates in the presence of both desferrioxamine (DFO) and ethylenediaminetetraacetic acid (EDTA). Abeta42 either alone or in the presence of Al(III) or Fe(III) formed beta-pleated sheets of plaque-like amyloids which were dissolved upon incubation with either chelator. Zn(II) inhibited whilst Cu(II) prevented the formation of beta-pleated sheets of Abeta42and neither of these influences were affected by incubation of the aged peptide aggregates with either DFO or EDTA. Freshly prepared solutions of Abeta42 either alone or in the presence of added Al(III) or Fe(III) did not form beta-pleated amyloid in the presence of DFO when incubated for up to 8 weeks. EDTA did not prevent beta-pleated amyloid formation in the same treatments and promoted beta-pleated amyloid formation in the presence of either Zn(II) or Cu(II). The presence of significant concentrations of Al(III) and Fe(III) as contaminants of 'Abeta42 only' preparations suggested that both of these metals were involved in either triggering the formation or stabilising the structure of beta-pleated amyloid. If the formation of such amyloid is critical to the aetiology of AD then the chelation of Al(III) and Fe(III) may prove to be a protective mechanism whilst the chelation of Cu(II) and Zn(II) without also chelating Al(III) and Fe(III) might actually exacerbate the condition.

As a nonessential element, aluminum is likely to be toxic both at low usual dietary levels in the long run (chronic toxicity) and at high therapeutic levels in shorter periods of time (acute toxicity). In both situations, aluminum toxicity is a direct function of aluminum bioavailability, which is itself dependent on Al(3+) solubility and charge neutralization. Dietary acids, by their intrinsic acidity and coordinating capacity, can extend the pH range, thus the section of the gastrointestinal tract, within which the Al(3+) ion remains soluble, and also help Al(3+) diffusion across the intestinal epithelium through the formation of neutral complex species. The present work examines the impact of glutamic acid, an essential amino acid also widely used in industrial food and drinks, on aluminum speciation in the gastrointestinal tract and blood plasma. Complex formation between the Al(3+) ion and glutamate has first been investigated through potentiometric titrations, complex stoichiometries being then checked by ESI mass spectrometry and NMR measurements. A series of mono- and polynuclear species has been characterized, whose influence on aluminum distribution in vivo has been assessed by computer simulation. The capacity of glutamate to maintain Al(3+) ions in solution under normal dietary conditions is predicted to be intermediate between glycine-like amino acids and succinate on the one hand, and tartrate and malate on the other hand, its Al(3+) neutralization effect being similar to that of succinate, tartrate and malate. These results, which point to a potential aggravating role of glutamate on aluminum gastrointestinal absorption, substantiate recent observations made on rats. In spite of the moderate effect expected from glutamate on aluminum bioavailability under most aluminum-based therapies investigated, attention is therefore called to the risk of glutamic acid ingestion simultaneously to any aluminum therapeutic form. Incidentally, the former implication of 'the' aluminum glutamate complex in the transfer of aluminum through the blood-brain barrier of aluminum loaded rats may effectively be attributed to one of the species characterized here, but is of no significance at all to aluminum contamination in humans, even at most extreme levels.

Aluminium toxicity may act in two distinct ways, depending on the level of contamination. Relatively low aluminium levels from environmental origin (mainly from drinking water poor in silica) have been shown to be statistically associated with senile dementias of Alzheimer type (chronic intoxication). In addition, high aluminium therapeutic levels (from phosphate binders, antacids, …) can induce different, more rapid, symptoms (acute intoxication). In all cases, aluminium toxicity is largely conditioned by aluminium bioavailability, which in turn hinges upon aluminium coordination chemistry in vivo. The highly polarising power of the Al3+ ion dictates its particular affinity for oxygen donors that abound in essential biomolecules and dietary substances. The influence of these substances on aluminium bioavailability, metabolism and toxicity can be assessed through animal models. However, understanding the mechanisms through which aluminium–ligand interactions may influence physiological processes on the molecular level requires a knowledge of the speciation of the metal in the main biofluids. Access to this critical information can a priori be gained through direct experimental analysis of relevant biological samples. It is in this way that aluminium protein-bound fractions, involving essentially transferrin, have been identified, but using such a direct approach to analyse the ultrafiltrable pool of the metal is a virtually insurmountable task, hence the necessity to have recourse to computer-aided speciation techniques based on simulation models. Following a previous review published in this journal on nearly the same topic [Coord. Chem. Rev. 149 (1996) 241], this article updates the knowledge available on both biological and chemical fronts. After a review of experimental investigations led on the roles of aluminium–ligand interactions in aluminium bioavailability, metabolism and toxicity, contributions of experimental and computer-aided speciation to the understanding of the relevant processes are then analysed. Significant progress has been made in the diverse aspects of the biological field, in particular, in relation to the role of dietary ligands on aluminium gastrointestinal absorption, excretion and tissue distribution. Also, very intensive research has been pursued on the design of new aluminium sequestering agents to treat acute intoxications. Some progress has also been made on the chemical side relative to computer-aided speciation applications to gastrointestinal and blood plasma conditions. However, the gap is increasing between the large body of observations made by physiologists and toxicologists and the few data painfully obtained by coordination chemists to interpret the relevant phenomena.

Aluminium bioavailability has raised much interest in the last two decades because of both the acute toxicity of high doses of the metal in some therapies and the chronic toxicity of dietary intakes in respect of senile dementias. In particular, a number of common dietary acids has recently been shown to play a part in this context. Essential amino acids, which are naturally released by digestion of vegetal and animal proteins, are also present as such in a large majority of industrial foods and drinks. They may, therefore, also interact with aluminium metabolism, especially in the gastrointestinal tract where they can reach quite appreciable levels. The objective of this work was to assess the extent of such potential interactions quantitatively. Aluminium complex formation equilibria of a series of essential amino acids involving glycine (used as a reference for all glycine-like amino acids), serine, threonine and histidine, were investigated under physiological conditions of ionic strength and temperature through glass electrode potentiometry and NMR. The only significant complex identified with the four ligands is M2LH−2. Simulations based on corresponding formation constants indicate that no particular risk is to be expected from these amino acids in normal conditions of alimentation and aluminium-based therapy.

As a nonessential element, aluminum may be toxic at both environmental and therapeutic levels, depending on ligand interactions. Dietary acids that normally occur in fruits and vegetables and commonly serve as taste enhancers are good ligands of the Al3+ ion. Malic acid is one of these and also one of the most predominant in food and beverages. The present paper reports an examination of its potential influence on aluminum bioavailability through speciation calculations based on Al(III)–malate complex formation constants especially determined for physiological conditions. According to the results obtained, malate appears to be extremely effective in maintaining Al(OH)3 soluble over the whole pH range of the small intestine under normal dietary conditions. In addition, two neutral Al(III)–malate complexes are formed whose percentages are maximum from very low malate levels. When aluminum is administered therapeutically as its trihydroxide, the amount of metal neutralized by malate peaks as its solubility pH range regresses to its original limits in the absence of malate. The enhancing effect of malate towards aluminum absorption is therefore virtually independent of the aluminum level in the gastrointestinal tract. The presence of phosphate in the gastrointestinal juice is expected to limit the potential influence of malate on aluminum absorption. Under normal dietary conditions, phosphate effectively reduces the fraction of aluminum neutralized by malate but without nullifying it. Aluminum phosphate is predicted to precipitate when aluminum levels are raised as with the administration of aluminum hydroxide, but a significant amount of neutral aluminum malate still remains in solution. Even therapeutic aluminum phosphate is not totally safe in the presence of malate, even at low malate concentrations. As plasma simulations predict that no compensatory effect in favor of aluminum excretion may be expected from malate, simultaneous ingestion of malic acid with any therapeutic aluminum salt should preferably be avoided.

The hypothesis is advanced that human betaA4--as opposed to rodent betaA4--may exert a protective function against the iron-induced oxidative stress associated with neurological diseases (notably Alzheimer's disease). Subsequent to its release by the host in response to oxidative injury, human betaA4 would interact with Cu(2+)ions whose level is correlatively elevated, adopting the 'aggregated' structure recently characterized by Atwood et al.(15). Then, depending on the oxidative state--hence the pH--of the medium, it might either return to its original structure if physiological pH is restored, or undergo site-specific copper-mediated oxidation and, finally, degradation. In this context, betaA4 pathogenicity could be due to an interfering mechanism preventing the degradation of the oxidized peptide, making its aggregation irreversible and inducing its final deposition. Coordination of side group oxygen donors of the oxidized peptide with 'hard' metal ions occurring in the physiological medium (notably Al(3+)) might be at the origin of this interference.

The capacity of a new ligand, 2-(hydroxyimino)propanohydroxamic acid (HPH, an analog of alanine), to bind the Al3+ ion has been examined. A particular objective was to check whether the HPH oximic group could take part in aluminum co-ordination together with the two hydroxamic oxygens, a possibility that could make HPH an efficient low-molecular-mass ligand for detrimental Al3+ ions in vivo. Glass electrode potentiometry, NMR spectroscopy and Electrospray Mass Spectrometry were used to investigate Al3+–HPH complex equilibria and their related structural aspects. The aluminium binding capacity of HPH was compared to that of other monohydroxamic acids. In spite of the lowest basicity of its hydroxamic moiety and the non-involvement of its oximic moiety in metal ion binding, HPH was found to be a very competitive ligand. Comparative quantum calculations performed on the systems involved suggest that electrostatic interactions induced by the ligand molecule are determining for complex stability.

Recent epidemiological studies have confirmed the existence of a correlation between aluminum level in low-silica drinking water and prevalence of Alzheimer's disease. Also, oral aluminum-based phosphate binders and antacids may induce acute aluminum toxicity. Whatever the source of the metal ingested, its bioavailability is a function of the chemical forms under which it occurs in the gastrointestinal tract, i.e. of the ligands with which the Al3+ ion may associate. Dietary acids in particular can favor the bioavailability of aluminum in different ways: by increasing its solubility, by complexing it into neutral species, and/or by acting indirectly on its absorption process. Among these, tartaric acid is commonly found in fruits and in industrial foods and drinks, and may therefore be ingested together with environmental or/and therapeutic aluminum. The present work examines its potential influence on aluminum bioavailability. Firstly, Al(III)-tartrate complex formation constants have been determined under physiological conditions (37 degrees C, 0.15 M NaCl). Then these constants have been used to simulate the influence of tartrate on aluminum speciation in different gastrointestinal situations in which phosphate was also taken into account. Under normal conditions of aluminum contamination, tartrate is expected to keep the metal soluble throughout the whole pH range of the small intestine, which is likely to enhance its bioavailability. Even at low concentrations, tartrate also gives rise to two neutral complexes that span over the 1.5-7.5 pH interval, a phenomenon that is aggravated by increased aluminum levels as may result from aluminum hydroxide therapy. The co-occurrence of dietary phosphate reduces the fraction of aluminum neutralized by tartrate under normal conditions, but this effect quickly decreases with increasing aluminum doses. Even the therapeutic use of aluminum phosphate is not expected to be totally safe in the presence of tartaric acid. As plasma simulations show that no aluminum mobilization can be expected from tartrate that could enhance aluminum excretion, avoiding ingestion of tartaric acid during any form of aluminum-based therapy appears advisable.

With a view to the possible use of copper(II)-*OH inactivating ligand (OIL) complexes as regulators of inflammation, the reactivity of the copper(II)-ascorbate system with hydrogen peroxide has been investigated in the presence of three key substances: histidine (the main copper(II) low molecular mass ligand in extracellular fluid), salicylic acid (the well-known nonsteroidal antiinflammatory drug, previously shown to be potentiated by copper(II) in animal models of inflammation), and anthranilic acid (an inactive substance by itself, known to be activated by copper(II) in the same models) at physiological pH (7.4) and inflammatory pH (5.5). Such substances may affect the amount of TBARS detected in solution following copper-mediated Fenton-like reactions through three distinct mechanisms: (i) by decreasing the Cu(II)/Cu(I) redox potential, i.e. at the expense of *OH radical production, (ii) by scavenging *OH radicals in the body of the solution, and/or (iii) by acting as a true OIL, i.e. at the expense of *OH detection. Redox potential measurements of initial solutions have been performed in parallel to TBARS determinations to help discriminate between different ligand influences. Computer-aided speciation has been used to understand the role of copper(II) distribution on the ligand effects characterised. Contrary to previous interpretations, histidine has been found to mainly affect *OH production by lowering the redox potential of the Cu(II)/Cu(I) couple. Salicylate, which has no effect on *OH production, has been confirmed to mainly scavenge *OH radicals in the body of the solution. Anthranilate, which both increases *OH production and decreases *OH detection, behaves as a potential OIL. These results tend to confirm our previous hypothesis that copper potentiation of antiinflammatory substances is indirect, i.e. independent of any interaction between metal and drug, whereas copper activation of substances that are inactive by themselves results from specific metal-substance interactions taking place at inflammatory sites.

While the involvement of environmental aluminum toxicity in the advent of senile dementias is still debated, acute aluminum toxicity of iatrogenic origin is well documented. So far, the only treatment available against it has been desferrioxamine (DFO), which induces major side effects. New drugs are thus highly desirable, and possible DFO substitutes have already been considered through various techniques. An important test for such new drugs is to assess their A1-mobilizing capacity in vivo. This can be done by computer-aided speciation provided formation constants for the corresponding A1(III) complexes are known beforehand. The present work reports an investigation of A1(III) complex equilibria with five sequestering ligands including DFO, and predicts the respective capacities of these to mobilize aluminum in vivo under normal and inflammatory conditions.

Previous in vivo studies devoted to the capacity of succinate to influence aluminum metabolism have led to apparent contradictory results. Understanding the mechanisms that lie behind such discrepancies requires a knowledge of aluminum–succinate interactions at the molecular level. In the absence of possible direct analysis of the ultrafiltrable fraction of aluminum in vivo, computer simulations can help quantify the mobilizing power of succinate towards aluminum in the main biofluids. Based on this technique, a first attempt to elucidate the above issue was made using especially determined aluminum–succinate formation constants. However, further investigations have led to reconsider the stoichiometry of the aluminum–succinate complexes characterized on that occasion. The present work deals with these new investigations. The results obtained confirm the great complexity of the aluminum–succinate system. No less than seven species, among which five polynuclear complexes, have been characterized in two series of independent experiments. New simulations indicate that succinate is expected to facilitate aluminum gastrointestinal absorption to a greater extent than initially predicted when the metal is administered as its trihydroxide, especially at high concentrations of the metal. In contrast, succinate is not able to significantly increase aluminum absorption when ingested concomitantly with aluminum phosphate. It is also confirmed that succinate cannot influence the fate of aluminum in blood plasma, which supports the view that the protective effect of succinate against aluminum toxicity in mice is not due to aluminum complexation.

Zinc is essential to the physiological activity of thymulin, a metallohormone involved in T-cell differentiation. The association of Zn2+ ions to the free nonapeptide (facteur thymique serique, or FTS) plays a highly specific role in thymulin activity, even though other metal ions, such as Cu2+, can partially restore it following incubation of the native hormone with Chelex 100. In order to throw light on the mechanisms through which metal ion-FTS association can influence thymulin activity, the present work reports an investigation into the stoichiometry and structure of copper(II) complexes with FTS under physiological conditions (37 degrees C, 0.15 mol dm(-3) NaCl). First, glass electrode potentiometry was used to study Cu(II)-FTS complex equilibria and to determine corresponding formation constants. Speciation calculations based on these constants were then used to design experiments with several spectroscopic techniques (electronic absorption, circular dichroism and EPR) which led to the determination of related structures. Except for the first bond to the Asn(9) C-terminal carboxylate group in the MLH species formed at low pH, copper coordination with FTS follows a rather unusual scheme. Classically, copper(II)-peptide bonding begins with the coordination of the Cu2+ aqua ion to an amino or an imidazole nitrogen that serves as an anchor site to initiate metal chelation through successive deprotonation of peptide bonds as the pH is raised. In the present case, the epsilon-terminal amino group of the Lys(3) residue, which binds copper in the relatively unstable ML complex, plays this initiating role. All the techniques used do effectively concur to demonstrate that, in spite of the 8-membered chelate ring required for multidentate coordination, the Lys3 amino group induces the chelation process that gives rise to the MLH-1 species, probably through the (Ala(2)-Lys(3)) site. Then, the classical scheme prevails with the deprotonation of the amide group at the Lys(3)-Ser(4) junction to form MLH-2. Finally, because of the important steric crowding around the metal ion which prevents it from binding to a third deprotonated peptide bond, MLH-3 forms via the deprotonation of the hydroxyl group of the Ser(4) residue. Implications of this particular coordination scheme for the discriminating role of the Lys3 residue versus thymulin activity are briefly discussed. (C) 1998 Elsevier Science S.A. All rights reserved.

It has long been established that copper complexes of inactive substances exert antiinflammatory activity and that copper complexes of nonsteroidal antiinflammatory drugs (NSAIDs) are more active than these drugs by themselves. Based on these observations, it was proposed that copper complexes of NAIDs are their active metabolites. This hypothesis was not confirmed for salicylic acid, however, as computer-aided speciation studies have shown that no copper-salicylate complex can reach significant levels in blood plasma. In view of this result, it was of interest to test with the same technique the influence on copper metabolism of an inactive substance known to be activated by copper. Anthranilic acid was chosen for this test in the present work. First, copper(II)-anthranilate interactions have been investigated by glass electrode potentiometry under physiological conditions. Given the key role of histidine as copper(II) ligand in blood plasma, copper(II)-histidine-anthranilate ternary equilibra have also been determined. Computer simulations of copper distribution have then been run relative to the two main biofluids in respect of global metabolism, i.e., gastrointestinal (g.i.) fluid and blood plasma. Like salicylic acid, anthranilic acid is expected to favor copper g.i. absorption, but cannot either exert any significant influence on plasma copper distribution. Clearly, the fact that anthranilate becomes antiinflammatory when administered with copper cannot originate in any effect of anthranilate on copper global metabolism. Speciation investigations have then been extended to the synovial fluid. Whereas salicylate does not appear to be a better ligand of copper in this medium than in blood plasma at any pH between 7.4 and 5.5, anthranilate on the contrary can mobilize increasing fractions of copper as the pH decreases, i.e., the more inflammation, the more copper is bound to anthranilate. This is in line with the recent observation that salicylate inactivates copper-induced .OH radicals through its bulk scavenging properties whereas .OH inactivation by anthranilate under the same conditions is a direct function of the copper-anthranilate binding. Anthranilate thus seems to correspond to the recently defined notion of .OH-inactivating ligand (OIL). More generally, these results provide a beginning of rationale for the antiinflammatory properties of copper complexes with substances that are active or inactive against inflammation by themselves. The extra antiinflammatory activity induced by copper on NSAIDs appears to be independent of any Cu(II)-NSAID association in vivo. On the contrary, the binding of inactive substances with copper(II) at inflammatory sites seems to be essential to their activation by copper.

Apparent contradictory results have been reported about the effect of succinate on aluminum toxicity, distribution and excretion in mice. Investigating the influence of this ligand on aluminum speciation in the main biofluids may help understand the above observations at the molecular level. In the absence of experimental access to ultrafiltrable aluminum speciation, computer simulations have been used in the gastrointestinal fluid and blood plasma, based on Al-succinate complex formation constants determined under physiological conditions. Calculations run for gastrointestinal conditions show that Al-succinate soluble complexes are formed in the 2 to 6 pH range-especially the neutral M(2)L(OH)(4)-which may enhance aluminum absorption. This influence, however, should be limited by dietary phosphate. In blood plasma, there is no possibility for succinate to mobilize a significant aluminum fraction, which confirms a recent suggestion that the possible protective effect of succinate against aluminum toxicity in mice may not be due to aluminum complexation.