Experimental aluminum pathology in rabbits: effects of hydrophilic and lipophilic compounds.

Aluminum lactate [Al(lact)3] (hydrophilic, hydrolytically unstable) and aluminum acetylacetonate [Al(acae)3] (lipophilic, hydrolytically stable) were tested as potential toxicants to rabbits upon IV administration both as aqueous solutions and as liposome suspensions. Both chemicals behaved as cardiotoxic agents when administered as aqueous solutions, but Al(acae)3 was at least two orders of magnitude more active than Al(lact)3. Al(acae)3, but not Al(lact)3, caused myocardial infarcts resembling those in humans (with contraction bands) at doses as low as 0.24 mg/kg body weight, as well as a prominent acanthocytosis. Al(lact)3, when administered as a liposome suspension, was about 300 times more toxic than in aqueous solution, although cardiac damage was not infarctual in character. Both chemical and physical speciation of aluminum(III) thus play an essential role in determining the toxicity of the metal. ImagesFIGURE 2.FIGURE 3.


Introduction
AluminumaII) is being actively investigated as a metal relevant to neurological disorders, such as Alzheimer's disease (1,2), dialysis dementia (3), osteomalacia (4), as well as being investigated for ecotoxicological effects (5). A possible role of aluminum uptake in the pathophysiology of dementias is suggested from recent epidemiological literature (7). However, from a biochemical point of view, the etiological connection between aluminum(llI) and such pathologies remains an open question (1).
In spite of toxicological (in vivo and in vitro) (6), enzymological (8), and biochemical (6,9) work over the last decade, information on the reactivity of aluminum(III) with Lewis bases suitable for mimicking possible interaction(s) of the metal with receptor sites of biomolecules is scarce (10,11). Moreover, while considerable information is available on the metallo-organic chemistry of aluminum(III) (12), the aqueous chemistry of the ion in the presence of organic ligands at neutral pH is relatively unexplored. Virtually all biological work, thus, depends on the use of aqueous solutions in which the chemical identity (the speciation) of the employed toxin is rather ill-defined. Three main types of aqueous aluminum have been used so far: a) salts of strong acids (e.g., chloride and sulfate); b) salts of moderately strong acid with appreciable metal bonding ability (e.g., lactate and tartrate); and c) hydrolytically stable neutral complexes, i.e., aluminum acetylacetonate [Al(acac)3] (acetylacetonate= 2,4-pentane-dionate) and aluminum maltolate [Al(malt)3] (maltolate = 3-hydroxy,2-methyl,4-pyronate).
Solutions of type a, once adjusted to pH 7.5, produce quantitatively Al(OH)3 (solubility equal to about 10-7 M).
Among solutions of type b, those obtained from aluminum lactate appear to be stable at pH 7.5, but they are, in fact, probably metastable (13). In this case, the actual identity of biologically active species is a matter of speculation (10). Solutions of type c are currently in use in our laboratory and in other laboratories (1,14); preliminary results obtained with Al(acac)3 (14) seem to indicate novel biological effects of this metal.
We report here on the first extensive toxicological work carried out with aluminum(llI) solutions in which the nature of the administered toxin is either well known, i.e., Al(acac)3, or at least can be reasonably proposed on the basis of preliminary new findings (15), i.e., aluminum lactate [Al(lact)3] (lactate=2-hydroxy-propanoate). Data referring to the use of liposomes as carriers of both Al(acac)3 and AI(lact)3 are also reported.

Animals and Chemicals
New Zealand white adult rabbits were obtained from Morini (Modena, Italy). Animals were maintained at controlled temperature and humidity and were supplied with water and standard rabbit food (Italiana Mangimi, Milano, Italy) ad libitum. The animals were anesthetized with Ketalar (Parke Davis) before being killed.
Al(lact)3, (Fluka) and DL-dipalmitoylphosphatidylcholine (DPPC) (Sigma) were of reageant grade and used without further purification. Al(acac)3 was prepared according to Yung and Reynolds (16) and was recrystallized from benzene/ light petroleum. The purity of Al(lact)3 and Al(acac)3 was checked by elemental analysis. Al(acac)3 was further controlled by 1H-NMR and IR spectrometry. Acetylacetone (2,4-pentanedione) (Janssen) was distilled before use. Ultrapure water was prepared by distillation in a fused silica subboiling still and stored in polyethylene bottles. Reagent-grade concentrated nitric acid (Prolabo) was distilled in a sub-boiling still and stored in polytetrfluoroethylene (PTFE) bottles. Diluted nitric acid (1 M) was prepared by weight in a PTFE bottle.
Preparation and Analysis of Liposome Suspensions of Al(acac)3 DPPC, 15 mg, and 105 mg of Al(acac)3 (about 0.32 mmole) were dissolved in 2.1 mL of ethanol (95%) at 70°C with stirring. The resulting solution was slowly injected by means of a 50-FL microsyringe into a sterile NaCl solution (9 g/L, 30 mL) at 55 to 60°C. Each 50-FL portion was injected in 10 to 15 sec; during injections, the solution was stirred and the syringe needle was kept about 1 cm below the surface. The solution was allowed to cool under stirring and excess DPPC and Al(acac)3 were removed by dialysis (cellulose membrane, 0 = 1 cm, average pore diameter = 24 A, exclusion limit 8000 to 15000 D, three treatments with 250 mL of physiological solution for 12 hr at 50 C). The final pH was about 7.0; the product had an opalescent appearance and was stored at 50C in glass vials. Monolayer and single compartment characteristics of the liposomes were checked by a turbidimetric method (17) and confirmed by electron transmission micrography (18). The liposomes were found to be stable for about 1 month (turbidimetric test) (17).
The total aluminum content of the liposome suspensions was determined as follows: 0.5 mL of the suspension plus 1.0 mL of water were acidified with 0.2 mL of concentrated HNO3 and kept at 800 C for 1 hr. The pH was adjusted at 7.4 by addition of 0.6 mL of aqueous NH3 (25% w/w) and of 5 mL of borate buffer (0.18 M boric acid, 5 mM sodium tetraborate, pH 7.4). Aluminum was then determined spectrophotometrically as the 8-hydroxyquinolinate complex after extraction with toluene (19).
The aluminum content of the liposomes was determined as follows: 0.5 mL aliquots of the suspension were centrifuged (15 min at 30,000 rpm) to remove the aqueous phase, and the organic residual was treated with 1.5 mL of water and 0.2 mL of concentrated HNO3; the resulting solution was analyzed as described above. The aluminum content of the liposomes was found to be about 50% of the total.
Preparation and Analysis of Liposome Suspensions of Al(lact)3 The procedure and molar amounts were the same as for Al(acac)3, with the exception that Al(lact)3 was dissolved in the physiological solution to give a 0.05 M concentration; the alcoholic solution contained only DPPC.
As most of the aluminum was expected to be in the external aqueous phase, the dialysis procedure was modified as follows. During dialysis, the external solution was continuously renewed by a peristaltic pump operating at about 35 mL/hr; the external solution was sampled at various times; and its aluminum content determined by ionic chromatography. A bell-shaped concentration versus time profile was obtained, with the aluminum concentration dropping to negligible values (0.4-0.5 /AM) after 5 days. The total aluminum content of the dialyzed suspensions was determined as described above; in view of the exhaustive nature of the dialysis procedure, the administered aluminum was assumed to be essentially all liposomes associated.

Administration Protocol, Chemoclinical Analyses, and Autoptical and Histological Evaluation
Animals were injected IV from the rostral auricolar vein under sterilite conditions. Solutions of AlOact)3 and AI(acac)3 were sterilized by autoclaving at 1100 C for 15 min. Control experiments were performed on three rabbits by injection of comparable amounts of free acetylacetone (aqueous solution, 300 Ag/day for 16 days).
A thorough autoptical examination was performed on most animals. All macroscopic lesions were noted and samples of brain, spinal cord, heart, lungs, liver, spleen, kidneys, and skeletal muscle were examined microscopically after 24 to 48 hr fixation in 10% buffered formalin. Sections were stained with hematoxylin-eosin; brain and spinal cord were also examined with Bodian's stain; heart was examined with the Alcian-PAS and van Gieson methods.

Decontamination of Analytical Equipment
All the vessels employed for aluminum determination in tissues were decontaminated by washing with a detergent and then with diluted nitric acid and were kept immersed in diluted nitric acid. Fused silica test tubes were boiled in concentrated nitric acid for 1 hr. All the above items were carefully rinsed with ultrapure water before use and allowed to dry in a dust-protected box. Micropipette tips were decon-taminated before use by loading twice with concentrated nitric acid and then with water.

Aluminum Determination in Tissues
Organs removed during the autoptic examination were stored at -20°C in decontaminated polystyrene vessels. A weighed tissue portion (2-5 g) was homogenized in a homemade PTFE potter (10 mL capacity) with 1 to 2 mL of water. The homogenate was then transferred to a decontaminated polypropylene test tube, freeze-dried, and stored at room temperature. The samples were dry ashed according to the following procedure (20): 150 to 300 mg of lyophilized material was directly weighed into a decontaminated 5-mL fused silica test tube and heated in a muffle furnace (temperature program: 1.5 hr each at 100, 150, 200, and 2500 C; overnight at 4800 C). After cooling, 1 mL of 1 M nitric acid was added, and the test tube was heated at approximately 1000C for 30 min. Occasionally, after this treatment, samples were found to contain small amounts of carbonaceous residue; in this case, the samples were treated in an ultrasonic bath for 1 hr and heated again as above. The samples were finally diluted with 1 mL of water and centrifuged. Aluminum was determined in the resulting clear solution by ion chromatography with colorimetric detection ( Fig. 1), using the operating conditions shown in Table 1.

Results and Discussion
Chemical Nature of Al(lact)3 and Al(acac)3 in Water Aluminum lactate has been frequently employed in toxicological research to prepare stock solutions of A3l ions. In  fact, this compound is reported to be "freely soluble in water" (21) and, unlike simple inorganic salts of All", does not give rise to aluminum hydroxide precipitation upon neutralization and dilution. This observation is in contradiction with thermodynamic expectations based on a recent potentiometric investigation of the AlF/H20/lactsystem (13). The apparent contradiction between thermodynamic prediction and experimental fact can be explained either by the existence, at pH 7, of other, yet unknown, coordination compounds under equilibrium conditions, or the operation of kinetic factors. This latter kind of behavior is not unexpected for All" complexes under nonacidic conditions (12). Indeed, no aluminum-lactate complexes exist in appreciable amount at pH 7.5; rather, the predominant species are aquohydroxo aluminum species, metastable toward Al(OH)3 formation (15). Based on available thermodynamic data (22), aqueous solutions of Al(acac)3 should be stable at the concentration levels employed in this work. This was confirmed by the stability with time of the electronic spectra of the solutions.
lbxicity Data A comparison of toxicity between aluminum lactate in aqueous solution (Al(lact)3, aq) and the same chemical carried by liposomes (Al(lact)3, ip) revealed that the latter is strongly (at least 300 times) more cardiotoxic than the former (Tables 2 and 3). For example, the data for animals 30 and 31 (Al(lact)3, p) and 18 and 19 (Al(lact)3, aq) show that similar CK figures were reached with analytical amounts of administered toxin that differ by about two orders of magnitude. This prominent enhancement of toxicity for Al' carried by the liposomes in their internal water micropools can be attributed to a biophysical speciation effect.
A liposome-based administration (likely to offer a more bioavailable form of Al"') allows for discrimination between specific cardiotoxic effects of the metal and the rather nonspecific ones caused by the same species at far greater concentrations to brain, liver, kidneys, and lungs ( Table 2). In particular, such effects appear to be prominent at Al(lact)3 doses greater than 70 mg/kg body weight. Under these conditions, such massive effects could have led to secondary myocardial damage (e.g., edema).
The biological effect of Al(acac)3,aq, estimated in terms of cardiotoxicity (Table 4), was also much higher (about 100 times) than that exhibited by Al(lact)3,aq, but the type of cardiac damage was quite different. Al(acac)3,aq produced a myocardial effect, with myocardial necrosis (infarct), prominent contraction bands, and multifocal myocarditis (the protonated free ligand, acetylacetone, is nontoxic). A 4-to 5-fold increase of aluminum concentration in the cardiac tissue paralleled the histopathological and chemoclinical data. Details of this experimentally induced pathology have been  'Numbers refer to pound registry. cTotal amount, including aqueous and liposome carried All" (see text). Number of days of treatment (d) and number of injections (i) in parentheses. dNumber of days after the last injection.
reported (14). It is worth noting that skeletal muscles undergo only indirect (secondary) damages.
The quite different biological effects on the same organ observed in the two cases clearly points to the importance of the speciation (i.e., the nature of the coordination sphere) of the administered metal. Cardiac muscle appears to be a target organ for aluminum (under pathologic conditions) for uremic patients (3). It is apparent that the chemical properties acquired by aluminum(III) in its peculiar speciation as Al(acac)3 (hydrolytic stability, resistance to scavenging, and lipophflicity) (24) facilitate the action of this artificial toxin not only in reaching the organ, but also in interfering with receptor(s) that are essential to myocardial cell function (14). The similarity of the cardiotoxic effect of Al(acac)3,aq and Al(acac)3,r, (Table 5) might seem surprising, in that a possible synergic action of the lipophilic vehicle with the lipophilic character of the coordination sphere could have been expected. Evidently, the above-mentioned properties of Al(acac)3 make this complex sufficiently well sulted in itself for selectively producing an infarctual cardiac lesion, such that the liposomal vehicle does not provide a significant synergic contribuution. The appreciable accumulation of  (14), the animal was monitored for an additional 50 days. Although CK and.LDH indexes returned to normal values after 3 days, histopathological analysis revealed that infarcted myocardial areas had undergone substitution of the affected fibers by collagen-rich scar tissue (Fig. 2). A similar pattern was observed also for rabbit 40, which died spontaneously 20 days after the end of treatment. Interestingly, microscopic (SEM) observations on blood samples from animals 39 and 40 seven days after termina- tion of treatment revealed a marked morphological anomaly of the erythrocytes (acanthocytosis) (Fig. 3). This anomaly, which was reported as associated with a form of encephalopathy in humans (25), may be interpreted as the consequence either of a damage to the emathopoietic system or of a direct action of All" on the erythrocyte membrane (9,26). An in vitro investigation appears to support this latter hypothesis (27).
As to the effect of Al(lact)3 on the central nervous system, this toxin was rather ineffective in producing appreciable damage when administered as an aqueous solution, in contrast with the observed conspicuous metal accumulation in the brain (Table 2). On the contrary, treatment with liposome-carried Al(lact)3 gave, in two cases (Table 3), posterior paraplegia with a wide infarcted area in the spinal cord. No neurological effects were observed with Al(acac)3. In general terms, the data reported here do not seem to indicate any connection between the IV injection of hydrolytically stable (lipophilic) Al(acac)3 or of hydrolytically (meta)stable (hydrophilic) Al(lact)3 and neurological disorders. The appparent contrast between our results and the neurological disorders with neurofilament degeneration observed upon SC administration of Al(lact)3 (28-30) may simply be the result of different administration protocols. In the case of Al(acac)3, the lack of effect may be due to the relatively short treatment times employed so far, together with an overwhelming cardiotoxic activity of the drug. It is possible that much lower doses and prolonged treatment times will lead to the production of neurologic effects.