J A Varner, K F Jensen, W Horvath and R L Isaacson
Binghamton, New York, and Research Triangle Park, North Carolina, USA

Abstracted from Brain Research 784 284-298 1998

The question addressed in this study was to understand why 0.5 ppm of aluminum fluoride in the drinking water of rats was associated with a larger increase in infections and mortality than with higher concentrations of 5 ppm and 50 ppm of aluminum fluoride in the drinking water. This higher mortality had been found in an earlier experiment when rats were given 0.5 ppm, 5 ppm or 50 ppm aluminum for 45 weeks in their drinking water.

The earlier experiment had been done to help determine whether aluminum in drinking water might have a role in aging-related neurological impairments. The suggestion had been made that when both fluoride and aluminum were present in drinking water they would form fluoroaluminum complexes, such as AlF3 in the stomach that, compared to their ionic forms, would be transported more easily into the blood stream and across the blood-brain barrier.

All the groups receiving AlF3 had shown the same significant reduction in the number of cells in certain areas of the hippocampus relative to the control group. The level of aluminum in the brains was almost double that of the control group, while morphological abnormalities were observed throughout the brain and its blood vessels. Dramatic behavioral differences were not seen.

One possible reason for the increased infections and mortality of animals in the 0.5 ppm AlF3 group was that the toxic agent of importance was the fluoride and that its absorption or its effects were antagonized by the higher levels of aluminum found with the higher doses of AlF3. The present study was undertaken to compare 0.5 ppm of aluminum fluoride, AlF3, with a comparable level of fluoride administered alone in the form of sodium fluoride.

Twenty seven adult Long-Evans rats were administered one of three treatments for 52 weeks. All groups received double distilled deionized drinking water. No additives were present in water for the control group. One treated group received water containing 0.5 ppm AlF3 (Al3+) giving a fluoride concentration of about 1 ppm. The other treated group had water containing 2.1 ppm of sodium fluoride, to provide an equivalent amount of fluoride (about 1 ppm) as for the AlF3 group.

Tissue aluminum levels of brain, liver, and kidney were assessed, and histological sections of brain were examined. For the morphological evaluations all counts and ratings were conducted by three individuals, two of whom were always blind to the treatment of the rats from whom the tissues came.

No differences were found between the body weights of rats in the different treatment groups although more rats died in the AlF3 group (5) than in the control group (1), P=0.05. A progressive decline in the appearance of the AlF3 animals was noted throughout the experiment, with the hair becoming sparse and the yellowing which occurs with age. The skin became dry, flaky and of a copper color. The aluminum levels in samples of brain were higher in both the AlF3 and NaF groups relative to the controls, P<0.01. The kidneys of the AlF3 group had higher aluminum levels compared to both the control and NaF groups, P<0.01. Liver aluminum levels did not differ between groups.

The effects of the two treatments on cerebrovascular and neuronal integrity were qualitatively and quantitatively different with the alterations being greater in animals in the AlF3 group than in the NaF group, and greater in the NaF group than in the controls. Examination of the distribution of the aluminum, using the Morin method for aluminum-fluorescence, showed the distribution in sections of brain to be exclusively associated with vasculature. This Al-fluorescence occurred in all three groups but there were consistent treatment-related differences in the intensity and extent of the aluminum-fluorescence. The NaF group had less Al-fluorescence associated with the vasculature than the control group, P<0.03. Renal changes were apparent in animals from both the NaF and AlF3 groups. More monocyte infiltration was present in the kidneys of the AlF3 group compared to the controls, P<0.05. No morphological abnormalities were present in the liver.

In the hippocampus, more moderately damaged and grossly abnormal cells were present in areas of the right hippocampus in the AlF3 group than in the control group, P<0.04, P<0.03. In neocortical layers 2 and 3 of the right hemisphere, the AlF3 group had more moderately damaged cells than the NaF group which had more normal cells, P<0.04. In neocortical layers 2 and 3 of the left hemisphere, the AlF3group had more abnormal cells than the controls, P<0.02. Corresponding increases in the frequency of normal appearing cells were observed in the neocortex of the control animals.

Neuronal density in the hippocampus was decreased in area CA3 in the left hemisphere of the AlF3 group compared to the controls, P<0.05. No differences were found in layers CA1 and CA4 in either hemisphere. Neuronal density was decreased in neocortical layers 2 and 3 in the left hemisphere of the AlF3 group compared to the controls and the NaF group, P<0.05. Neuronal density was also reduced in layers 5 and 6 of the left hemisphere of the AlF3 group compared to the controls, P<0.03, P<0.03. No differences in neuronal density were found, in any of the groups, in layer 4 of the left hemisphere or in any of the layers of the cortex in the right hemisphere.

While in control animals the localization of IgM was largely restricted to the vascular lumen, indicating the integrity of the blood-brain barrier to IgM, in the AlF3 and NaF groups staining for IgM in the neural parenchyma was increased in the right hemisphere, P<0.03, P<0.01. No differences were found among the groups in the left hemisphere. Minor amounts of IgM immunostaining were detectable in the hippocampus and dentate gyrus but no significant differences were present.

Differences in the amount of immunoreactivity for beta-amyloid relative to that for amyloid A were most prominent in the vasculature of the dorsal thalamus. The control group had few instances of immunoreactivity while the tissue of animals from the AlF3 group demonstrated a bimodal distribution of reaction product with either no reaction product staining or a high level. The AlF3 group had more immunoreactivity for beta-amyloid in the lateral posterior thalamic areas of both hemispheres relative to the controls, left P<0.05, right P<0.01. The NaF group differed from the control for immunoreactivity for beta-amyloid in the right lateral posterior thalamic area with the controls having low reactivity and the NaF group having no or high levels of immunoreactivity, P<0.01.

Discussion and Conclusions
The high mortality rate in the animals receiving 0.5 ppm of aluminum in their drinking water found in the first study was replicated in this second study. Since the administration of sodium fluoride alone did not produce a similar mortality rate, this effect does not appear to be directly related to fluoride intake. The appearance of the AlF3 animals, with sparse hair and a copper-colored hypermelanosis of the underlying skin, may be indicative of several diseases including chronic renal failure. Histological evidence of glomerular distortions was present in both the AlF3 and NaF groups.

The overall aluminum level of the kidneys in the AlF3 group was nearly twice that of the NaF group and possibly may have been associated with a greater level of physiological dysfunction. The kidney is critical to the elimination of both aluminum and fluoride, and these changes may have impaired the elimination of these elements, detoxification in general or homeostasis of other ions such as calcium.

Both the AlF3 and NaF groups had increased brain aluminum levels relative to controls. The aluminum level in the NaF group was double that of controls and the aluminum level of the AlF3 group even greater. The aluminum detected in the controls and NaF groups is most likely to have come from the rat chow where the reported levels of aluminum range from 150 ppm to 8300 ppm. Fluoride commonly occurs in food and water and is almost completely and quickly absorbed from the gastrointestinal tract. In the present experiment, the AlF3 in the drinking water was prepared to form optimally a fluoroaluminum species capable of crossing the gut and vascular barriers. It is possible that the sodium fluoride-treated group was able to form some amount of an AlF3 also capable of becoming bioavailable.

In general, the reduction of neuronal density in the neocortex of the left hemisphere was more prominent in the AlF3 group than the NaF and control groups. Cellular abnormalities in the form of chromatin clumping, enhanced protein staining, pyknosis, vacuolation, and the presence of ghost-like cells were also more common in the AlF3 group in the left hemisphere. Vascular Al-fluorescence was more pronounced throughout most structures of the left hemisphere of the AlF3 group. The hippocampus was an exception to these findings with abnormalities being found only in the right hemisphere in the CA1 and CA4 areas of both the AlF3 and NaF groups. The right hippocampus also had higher levels of aluminum-induced fluorescence than the left hippocampus.

Possible cellular mechanisms which might underlie the association between regional aluminum accumulation and regional patterns of neuronal injury include transferrin transport, calcium homeostasis and second messenger systems, alterations in neuronal cytoskeleton, and alterations in the cerebrovasculature. Transferrin transport carries certain metals, including aluminum, into the cells by a receptor complex located on plasma membranes. Neurotoxic reactions to aluminum may also result from disruption of ion homeostasis and second messenger systems, in particular G-proteins. The activation of G-proteins in turn initiates a chain of reactions beginning with adenylate cyclase, cAMP, and protein kinases that ultimately result in increased phosphorylation of various substrates. G-proteins are involved in the regulation of ion channels, metabolism, gene expression, and cytoskeletal structures via second messenger systems.

The deposition of aluminum in the cerebrovasculature was seen to have the potential also to contribute to the observed neuronal injury by altering the blood-brain barrier or cerebral blood flow. Both the AlF3and NaF groups had increased levels of IgM in the cortex of the right hemisphere compared to the controls. Serum proteins, such as IgM and other antibodies, are typically excluded from neuronal tissue by the blood-brain barrier. The increased immunoreactivity of IgM in the present study may be indicative of a compromise in the blood-brain barrier occurring possibly through changes in lipophilicity or the potentiation of existing transport mechanisms.

Striking parallels were seen between aluminum-induced alterations in cerebrovasculature and those associated with Alzheimer’s disease and other forms of dementia where microvascular abnormalities display a regional and laminar specificity in accordance with neuronal degeneration patterns, suggesting that the alterations in the cerebrovasculature may be a primary event in neurodegenerative disease. Accumulation of beta-amyloid has been found in the cerebrovasculature and spinal cord vasculature of elderly persons and may be a consequence of the development of blood vessel abnormalities. Vascular tissues produce beta-amyloid and the vascular deposition of beta-amyloid may result from alterations in the basement membrane of blood vessels that frequently occurs with ageing. Although it is uncertain whether beta-amyloid is a causative factor in neurodegeneration or a secondary consequence of injury, aluminum appears to have the capacity to influence the distribution and properties of beta-amyloid. In the present study animals in both the AlF3 and NaF groups exhibited a bimodal distribution of vascular beta-amyloid in both hemispheres in the lateral posterior thalamus being either absent or present in high levels in both groups. This increase in vascular beta-amyloid in the lateral posterior thalamus may be causally related to the neuronal degeneration found in the area to which it is principally connected, the superior parietal cortex.

In contrast, the control group consistently had only low levels of immunostaining for beta-amyloid, a unimodal distribution. While the presence of low levels of beta-amyloid in the controls was understandable as being the possible result of a natural continuing, but limited, production of the protein throughout life, the bimodal distribution of beta-amyloid in the AlF3 and NaF groups was not understood.

While the present results do not address the causal mechanisms of aluminum-induced neural degeneration, they do demonstrate that ingested aluminum reaches the brain, and in such animals there exists evidence of neural injury. While the small amount of aluminum of 0.5 ppm AlF3 in the drinking water of rats required for neurotoxic effects was seen as surprising, the neurotoxic results of NaF at the dose given in the present study, 2.1 ppm or about 1 ppm of fluoride, was seen as even more so.

As dietary sources of fluoride are additive in animals it was considered that, for the animals in the NaF group, the fluoride in rat chow, such as Purina Rodent Laboratory Chow, together with their drinking exceeded tolerable levels. Fluoride has diverse actions on a variety of cellular and physiological functions including the inhibition of a variety of enzymes, a corrosive action in acid media, and the production of hypocalcemia, hyperkalemia and possibly cerebral impairment.

Few chronic toxicity studies of fluoride have included extensive histological characterization of injury to the brain but have usually been limited to weight loss, dental and skeletal changes, indicators of carcinogenesis, and damage to soft tissues. The results of the present study indicate that more intensive neuropathological evaluations of fluoride effects on the brain is likely to be of value.

In summary, the chronic administration of aluminum fluoride and sodium fluoride in the drinking water of rats resulted in distinct morphological alterations in the brain, including effects on neurones and the cerebrovasculature. Further studies of aluminum fluoride and sodium fluoride are needed to establish the relative importance of a variety of potential mechanisms contributing to the observed effects as well as to determine the potential involvement of these agents in neurogenerative diseases.

Key words: Aluminum-fluoride; Amyloid; Brain; Cerebrovasculature; Hippocampus; Neurotoxicity; Rat; Sodium fluoride.
Reprints: Julie A Varner, Lineberry Research Associates, PO Box 14626, Research Triangle Park, North Carolina 27709, USA. Fax: +1-919-547-0974.