- Email: [email protected]
The puzzle of Alzheimer's disease (AD)
Medical Hypotheses I
I Matid tfypoihma (1992) 38.5-10 0L4mgmnhpuKLtd1992
The Puzzle of Alzheimer’s Disease (AD) H. D. LEHMANN NM Natutwissenschaftliches und Medizinisches lnstitut an der Universit& Tiibingen in Reutlingen, Gustav- Werner-Str 3, D-74 10 Reutlingen, Germany
Abstract-A compromised defensive system of brain cells against aluminium, together with local defects in glucose metabolism, causes AD. Lack of citrate is a driving force and free cis-aconitate or glutamate are potential carriers, which enable the exotoxin to cross lipid membranes. Only a few aluminium ions replace magnesium in key positions. They block the reversibility of phosphorylation reactions, which are important for short term memory: sensitization of the insulin receptor and protein phosphorylations. Due to disturbed phosphorylation of the cytoskeleton, protein synthesis runs out of balance. Efforts to restore the disturbed reactions result in AD specific deposits. Aluminium ions are the common cause for the induction of AD pathogenesis in patients with genetic defects, with mechanical brain lesions or with minor infarcts, as well as with changes in the relation between numbers of neurons and neuron nursing glia cells due to age.
Introduction
have shown AD to be a disease not restricted only to the brain. Together with a discussion of the role of magnesium (6), this may renew interest in the Al hypothesis. There is a possibility that this cation may be the missing link between the different AD research areas, as well as between different types of dementia. The following thoughts are deduced from recent AD literature, as well as from fundamental biochemistry. They have the aim of bringing together the separate AD research areas by a modified aluminium hypothesis and of helping to understand differences or similarities of AD to other dementia.
The study of Alzheimer’s Disease (AD) has become highly specialised. There is a wide variety of different areas of research: the molecular biology of the amyloid precursor protein and how its fragments agglomerate; the deficiencies in brain glucose metabolism; neurotransmitter deficiencies; the research on alterations in neural cytoskeletal proteins, as well as on differences found in fibroblast cell cultures. One hypothesis, discussed controversially some years ago, proposed a role for the exotoxin aluminium (Al) in AD (1, 2), which has neither been proven nor disproven. A recent survey on the association between Alcontaining products and AD stated ‘provocative but inconclusive results’ (3). Some recent publications on transferrin deficienties in AD (4, 5) and on work with non-neural ceils, Date received 14 March 1991 Date accepted 13 November 1991
The protection of brain cells against the exotoxin aluminium
Aluminium is ubiquitous on earth. With food and beverage we all ingest several milligrams of Al every
5
6
day; rather larger amounts are taken in by patients using Al containing antacida. Al is an element with no known biological function and has some well-known toxic properties (7). Why are there no problems with Al normally? How does the organism overcome the permanent challenge of incoming toxic Al? To protect against aluminium toxicity, the organism has developed a system of barriers, to efficiently eliminate and exclude Al. The small AlJ+-cation, with a high charge does not exist in a free form under physiological conditions. It is complexed and has a high affinity to some ligands. Normally, Al is detoxified by forming a complex. There is a distinct hierarchy in the affinity of Al to substances occurring in a biological milieu. Phosphates and carboxylates are the preferred ligands, and the molecules ferritin, transferrin and citrate show the highest affinity and arc used by the organism to detoxify Al. There are three barriers which prevent dietary Alions from intrusion into brain cells: the intestine, the blood-brain barrier and the cell membrane. Most Al ingested leaves the intestines without resorption. Al intruding cells in the intestinal wall is immobilized by ferritin. It leaves the body with the faeces when the cells are peeled off the intestinal wall. Only a small percentage of Al reaches the blood stream. There it is bound to transferrin @KA 14.7) and transported for renal elimination. Under normal conditions the Al-concentration in the blood is, therefore, very low (6ng/ml). Transfer-i-in,the transport vehicle of iron, is able to pass cell membranes by endocytosis at sites where special receptors are available. In the rat, high numbers of these receptors exist in those brain regions which are found sensitive to AD in man (3). Bound to transferrin, Al can pass the blood-brain-barrier by transcytosis. Banks and Kastin (8) report on damages of this barrier due to Al. They found 35 rig/ml Al in the liquor, i.e. an increase of Al behind the bloodbrain-barrier in normal, not demented patients. This reflects the rather low permeability of the cerebrovascular endothelium with its tight junctions. Some Al released from transferrin in the brain may be detoxified by ferritin (9). If there is not sufficient binding capacity of transferrin and ferritin, the next candidate in the affinity hierarchy for complexing with Al is citrate. Its affinity for Al is less than that for transfenin (pKA 14.2). In the system citrate/Al3+/H+ the complex is charged negatively at physiological pH (10). This charge prevents from passing across lipid cell membranes. The citrate concentration in the body fluids is normally high enough to keep Al in a detoxified state, even with high loads, as found in dialysis patients. In
MEDICALHYIWI-HESES
blood, the citrate concentration is about 10-4 mol/l. Citrate is an intermediate of the glucose oxidation process which takes place in the cellular mitochondria (Krebs or citrate cycle). The citrate concentration is higher intracellularly than extracellulary resulting in a gradient over the cell membrane. A citrate leakage out of the cell may be an additional protection for the cell against the intrusion of Al-citrate complexes. The role of Al in AD
Al has long been suspected of playing a role in AD, but the mean concentration levels in brain tissue are similar in AD patients and normal controls (11). Comparable results are also found for serum and liquor. Al-analysis of Alzheimer brains is difficult, either because of the appearance of Al in small focal sites or in very low concentrations. Recent investigations found no accumulation. Most reports in the literature do not state in which complex form the Al is found. Only within affected neurons of Alzheimer patients have high concentrations of Al been found by several laboratories. This result can be understood if both a carrier and a driving force for Al in complexed form transport the ion over the lipid membrane into the intracellular region, i.e. against the normal citrate concentration gradient. This would lead to an intracellular accumulation of Al, when the ligand is metaboliscd within the cell. These conditions are found in several dementia: in AD as well as in Multi-Infarct-Dementia (MID). A local small infarct causes hypoxia and glucose deficiency, which lowers the pH locally. This in turn alters the equilibrium between citrate and its successor products in the Krebs cycle. Citrate is normally dehydrated to cis-aconitate and rehydrated to iso-citrate by aconitase, in a state bound to the enzymes in the mitochondria. There is usually no leakage of cis-aconitate to the cell membrane, as its concentration is too low and the rehydration is rapid at normal pH conditions. In the cytosol, there is an iso-enzyme of isocitrate-dehydrogenase, which splits iso-citrate to CO, and a-ketoglutarate. The Krebs cycle enzyme activity in the mitochondria depends on the cofactor NAD+, the cytosolic iso-enzyme on NADP+. Insufficient NADP+ supply in the cytosol may result in an escape of cis-aconitate under these conditions Cisaconitate escaped from one cell may play a detrimental role as Al-carrier for cells in the neighbourhood. Cis-aconitate probably has a high affinity for Al. This ligand is an appropriate Al-carrier for its transport across the cell membrane, because the complex is not charged and highly lipophilic, i.e. soluble in
7
THE PUZZLE OF AIZHEIMER’S DISEASE
the lipid cell membrane. Cis-aconitate, therefore, is thought to be the pathological Al-carrier in AD and MID. It is generated at low pH values and reaches the cell membrane when the enzyme is defective or its cofactor NADP+ is missing. Guy et al (12) used human neuro-blastoma cells as an in vitro model to show the different carrier activities of citrate, EDTA and malt01 for Al. Rather low values were found for citrate. Another potential carrier for Al may be glutamate, as proposed by Deloncle et al (13). In AD brains, glutamate levels are comparable to normal controls. Glutamate can be used in the Krebs cycle as a substrate for energy gain. Its neurotransmiuer function, together with the normal or even lowered (14) concentrations found in the cerebrospinal fluid of AD patients, make it difficult to understand its detrimental role as an Al-carrier in restricted brain areas. On the other hand, due to its excitotoxic properties, fatal consequences of glutamate on sensitive neurons are known. The energy demand of cells near an infarct-r endothelium cells, neurons or glia cells-will establish a driving force for extracellular metabolizable substrates to enter the cell. Within the cell, the potential Al-carriers cis-aconitate or glutamate are metabolized in the Krebs cycle, if the Al finds another ligand of high affinity. What are the consequences?
In close neighbourhood to the infarct, different cell types become damaged. Within the affinity hierarchy of ligands, Al favours next phosphate groups. Phosphate is present in biological media in rather high amounts, but only certain phosphate assemblies are supposed to have a very high affinity for Al as for instance ATP or inositol 1, 4, S-triphosphate (15). The appropriate geometry within connected groups is given by the dimensions of the cation. AP+ is similar in size to Mg2+. The ion radius for Al3+ with coordination number 6 is 54 pm, the value for Mg2+ with coordination number 4 is 57 pm. (These figures are very similar to the ion radius of Ni2+ in a tetragonal configuration: 55 pm. On Ni-rich soils certain plants detoxify and accumulate this element in high amounts in form of a negatively charged citrate complex). In the absence of citrate, the trivalent Al will complex to phosphate groups which are normally paired with the bivalent cation, Mg2+. This is fatal, because it blocks several vital phosphorylation reactions by inhibiting their reversibility. Reversible phosphorylations need bivalent cations as co-factors having a coordination number of 4, two sites of the maximum coordination
number being left open. This is possible for bivalent Mg but not for trivalent Al with its preferred coordination number of 6. Al with its high affinity, thus, blocks the reversible transfer of phosphate groups. A multitude of fatal consequences results: A.
The glucose supply is stopped as a consequence of the de-activation of the insulin receptor, which is sensitized to glucose by a Mg/Mn-dependent kinase phosphorylation. Reversible tyrosine phosphorylation is important for the function of the insulin receptor. In AD, an altered protein tyrosine phosphorylation is reported (16). Low insulin receptor sensitivity multiplies the effect of a small infarct. It enforces the influx of substances other than glucose to keep the Krebs cycle running, including Al-complexed cis-aconitate and glutamate. Glutamate metabolism is accompanied by transamination reactions. Ammonia is generated endogenously in the brain of Alzheimer patients (17). Recent amino acid investigations indicate selective degeneration of glutamate-sensitive neurons (18, 19). The concentrations of amino-acids occurring in liquor were found to be identical for MID and AD (20).
B.
Reversible phosphorylations of neuronal cytoskeletal proteins will be blocked (21, 22, 23). MAP kinase activated by nerve growth factor is dependant on tyrosine phosphorylation, too. This has been shown on PC 12 cells (24). Differences found in cytoskeletal structures in fibroblasts from Alzheimer patients, compared to controls, are caused by different phosphorylation states (25). In neurons, the phosphorylation of cytoskeletal proteins controls dendritic growth. The reversibility of reactions may be important, especially for the short time memory: information brought via neurotransmitters is transformed by phosphorylation reactions. Irreversibility of this process will result in uncontrolled dendritic growth and also in an overreaction of the proteolytic activities, directed against dendritic growth. This, further, will induce the overreaction of the porteolysis-controlling system, i.e. the synthesis of the amyloid precursor protein by microglia (26). The effect of protein phosphoryiation on amyloid precursor protein generation has been shown (27). Furthermore, Al-induced changes in neuronal chromatin are thought to reduce gene transcription (28, 29). This overreaction of the control system has detrimental consequences. In areas with a local energy deficit, the disintegration of the protease in-
MEDICAL HYPOTHESES
hibition system is incomplete. Ubiquitine, found in AD deposits is a marker substance for incomplete proteolysis. The agglomerating pieces of the precursor protein, together with tau, form the neurofibrils specific for AD. The amyloid is found around neurons with wild dendritic growth, in synapses and within brain blood vessels. Both forms of deposits typical for AD have secondary consequences: Together with the folloing point (C), they induce the psychological manifestations of the disease. C.
Several other phosphorylation reactions are disturbed. The production of neurotransmitters (DOPA; dopamin, serotonin) will be blocked due to the inhibition of the dihydropteridine reductase (30). The inhibition of choline-actyltransferase also results in neurotransmitter deficiencies.
The very slow continuous intrusion of only a few Alions induces a permanent stress situation (31), if the are not detoxified by ferritin (32). Neuronal death due to the neuroexcitatory glutamate is a fatal secondary consequence, as well as the amyloid deposits (33) and the disturbances in neurotransmitter synthesis. According to this modified aluminium hypothesis, three factors create together the conditions for AD on a primary level by blocking the reversibility of the fundamental processes necessary for short term memory. 1.
2.
3.
A malfunction of Al elimination and its detoxifying system: a qualitative or quantitative acquired or inherited defect of transferrin. Removal of Al by appropriate chelating agents, e.g. by deferoxamine (34, 35) seems, therefore, to be a reasonable therapy for AD. Small infarcts result in a local hypoxia, acidosis and a deficiency of glucose. The Krebs cycle supplies become insufficient. Extracellular substrates are used to compensate for the deficiency. Due to a local deficiency of NADPt, cis-aconitate may reach the lipid cell membranes. This may occur in glia cells, which are afflicted by Al, as well as in neurons or endothelial cells and may influence neurons in the neighbourhood.
Can the hypothesis explain special features of AD?
Pathological observations of AD concentrate on the amyloid deposits in the brain. A hypothesis must explain why different cell types are afflicted in focuses
in these areas, although recent observation on cells of non-neuronal origin from Alzheimer patients have shown AD to be a systemic illness, not restricted to the brain (36). There are several possibilities: 1.
2.
3.
In the chicken model, it has been shown (37) that glucose uptake and a following presynaptic protein phosphorylation in distinct brain regions are prerequisites for short term memory. In the evolutionary old part of the brain, the ratio of supplying cells per neuron may be lower than in more recently evolved areas. With age, this ratio is lowered by naturally occurring cell death. It is possible that here a ratio may he reached that is insufficient for normal function. For a slowly proceeding illness, this is, of course, only of interest for cells in the neighbourhood of surviving neurons. This is different to acute brain damage due to neuronal death. Another possibility may lie in the different number of transferrin receptors shown in the rat brain (4).
Another specific problem is that AD is found in two forms: it occurs as a hereditary illness, as well as spontaneously. AD is found as a genetic defect either in combination with Down’s syndrome or occurring in adults (FAD = familial AD). On the other hand, AD occurs in a spontaneous form in the elderly. There is a statistical correlation between the increase in age of our population and the number of AD patients. A high Al load alone, as given in many dialysis patients, is not sufficient to produce AD. Very often AD and MID are combined (38). Mechanical brain lesions are a risk factor for AD (‘dementia pugilistica’). The table shows a potential link between different dementiae. It tries to explain the common appearance of AD in patients with different histories. Although many questions remain unanswered, some controversial features typical for AD can be explained by a hypothesis that proposes that the illness results from an overlap of several causes, linked by Al. The normal protection system of the organism, on the other hand, does not justify a fear of Al nor provide hope for clear results from epidemiologic studies on Al in AD. Conclusion Very small amounts of Al in serum or liquor-amounts not easy to analyze and which do not exceed normal values-are fatal in combination with factors introduced by small infarcts and by age statistics. They block the reversibility of vital phosphorylation reactions if they gain access to the interior of cells. Symptoms of AD can be related directly or indirectly to
THE PUZZU
9
OF AIZHEIMER’S DISEASE
Table Features common
to different dementiae.
Are there Iinks?
some Normal
Dialysis patients
Mulfiinfirctalion
?
n0
Systemically compromised Al-elimination
“0
Locally compromized Krebs Cycle (Glucose and 02-Deficiency)
IlO
no
?
Locally insufficient low relation between glia cells and neurons
no
no
?
High aluminium
no
Resulr
load
“0
In conse-
dialysir patients
?
yes
lowering
Adults with Down Syndmne
with
normal
yes dialysis dementia
“0
MID
the blockage of reversible processes by Al, which replaces Mg-ions at key positions. References 1. Crapper Mclachlan DR. Aluminium and Alzheimer’s disease Neurobiology of Aging 7: 532, 1986. 2. Shore D. Wyatt R 1. Ahtminium and Alzheimer’s Disease. J New Mental Disease 171: 553, 1983. 3. Graves A B et al. The association between aluminium-containing products and Alzheimer’s Disease. J Clin Epidemiol 43: 35, 1990. 4. Pullen R G L et al. Gallitmt-67 as a potential marker for aluminium transpott in rat brain. Implications for Alzheimer’s Disease. J of Neurochemistry 55: 1990. Farrar G et al. Defective gallium-transfertin binding in Alzheimer’s disease and Down Syndrome: possible mechanism for accumulation of aluminium in brain. The Lancet 335: 747, 1990. 5. Connor J R. Iron storage and transport proteins in the brain in aging and Alzheimer’s Disease (abstract). Neurobiology of Aging 11: 77. 1990; Camor J R et al. Cellular distribution of transferrin, ferritin and iron in normal and aged human brains. J Neurosci Res 27:595, 1990. 6. Glick J L. Dcmentias: the role of magnesium deficiency and an hypothesis concerning the patbogenesis of Alzbeimer’s Disease. Medical Hypotheses 31: 211, 1990. 7. Ganrot P 0. Metabolism and possible health effects of aluminium. Enviratmental Health Perspectives 65: 363, 1986. 8. Banks W A, Kastin A J. Aluminium induced neurotoxicity: Alterations in membrane functions at the blood-brain-barrier. Neuroscience and Biobehavioral Reviews 13: 47, 1989. 9. Fleming J. Joshi J G. Ferritin: Isolation of aluminium-ferritin complexes from brain. Biochemistry 84: 7866. 1987. 10. Ghmann L 0. Equilibrium and structural studies of Si IV and Al III in aqueous solution. 17. Stable and metastable complexes in the system H+-A13+-citric acid. Inorg Chem 27: 2565. 1988. 11. Jacobs R W et al. A Reexamination of aluminium in Alzbeimer’s Disease: Analysis by energy dispersive X-ray microprobe and flameless atcmic absorption spectrophotometry. Canad J of Neural Sci 16: 498, 1989.
?
?
In old paiients ?
(inherited)
(inherited)
?
? (due to infarcts)
yes
(yes)
?
? (inherited)
lowering with age
lowering with age
age
yes
guence lo In JVXUl~ head traumata patients
no
I-IO
no
“0
FAD + Down Syndrome
FAD
AD
AD
12. Guy S P et al. Uptake of aluminium by human neuroblastoma cells. Biochem Sot Transactions 18: 392. 1990. 13. Deloncie R et al. Alzheimer’s disease and dementia syndromes consecutive to inbalanced mineral metabolisms subsequent to blood brain barrier alteration (abstract) 1. Symposium at metal ions in biology and medicine, Reims 16th May 1990. in cerebrospinal 14. Basun H et al. Amino acid concentrations fluid and plasma in Alzbeimer’s Disease and healthy control subjects. J Neural Transm 2: 295, 1990. 15. Birchall J D, ChappeII J S. AIuminium. chemical physiology, and Alzheimer Disease. The Lancet 1008, 1988. 16. Shapiro I P et al. Altered protein pbosphotylation in Alxheimer’s Disease. J Neurochem 56: 1154. 1991. 17. Hoyer S et al. Ammonia is endogenously generated in the brain in the presence of presumed and verified dementia of Alzheimer type. Neuroscience Letters 117: 358, 1990. 18. Choi D W. Glutamate neurotoxicity and diseases of the nervous system (Review). Neuron 1: 623, 1988. 19. Lowe S L et al. Ante mortem cerebral amino acid concentrations indicate selective degeration of glutamate-enriched neurons in Alzheimer’s Disease. Nemo Science 38: 571, 1990. 20. Degrell J et al. Amino acid concentrations in cerebrospinal fluid in presenile and senile dementia of Alzheimer type and multi-infarct dementia. Arch Gerontol Geriatr 9: 123. 1989. 21. Blass J P et al. Induction of Alzheimer antigens by an uncoupler of oxidative phosphotylation. Archiv Neural 47: 864, 1990. 22. van Huynh T et al. Reduced protein kinase C immunoreactivity and altered protein phosphorylatiat in Alzheimer’s Disease fib&lasts. Arch Neurol 46: 1195, 1989. 23. Steiner B et al. Phosphorylation of microtubule-associated protein tau: identification of the site for Caz+-calmodulin dependent kinase and relationship with tau phosphorylation in Alzheimer tangles. EMBO J 9: 3539. 1990. 24. Ganez N. Cohen P Dissection of protein kinase cascade by which nerve growth factor activates MAP kinases. Nature 353: 170, 1991. 25. Takeda M et al. Changes in adhesion efficiency and vimetttin distribution of fibroblasts from familial Alzheimer’s Disease patients. Acta Neurologica Scandinavica 82: 238, 1990.
10 26.
27.
28.
29.
30.
31.
32.
MEDICAL HYKYlXE%8
Perlmutter L S et al. Morphologic association between microglia and senile plaque amyloid in Alxheimer’s Disease.. Neuroscience Letters 119: 32. 1990. Buxbaum J D et al. Processing of Alzheimer B/A4 amyloid precursor protein: modulation by agents that regulate protein phosphotylation. Proc NatJ Acad Sci USA. 87: 6003, 1990. Lukiw W J, Crapper McLachlan D R. Chromatin structure and gene expression in Alrheimer’s Disease. Mol Brain Res 7: 227. 1990. Crapper McLachlan D R et al. New evidnece for an active role of aluminium in Alzheimer’s Disease. Canad Journal of Neurol Sci Suppl4, 16: 490, 1989. Ahmann P et al. Serum aluminium levels and erythrocyte dihydropteridine reductase activity in patients on hernodialysis. New England J Med 317: 80, 1987. Papasozomenos S C, Su Y. Altered phosphorylation of Tprotein in heat-shocked rats and patients with Alzheimer’s Disease. Proc Natl Acad Sci USA 88: 4543, 1991. Grundke-Iqbal I et al. Ferritin is a component of the neuritic
(senile) plaque in Alrheimer dementia. Acta Neuropathol81: 105, 1990. 33. Beyreuther K et al. Demenz von Alzheimer Typ. Deutsche Apotheker Zeitung 131: 1414. 1991. 34. Kruck T P A et al. Suppression of deteroxamine mesylate treatment-induced side effects by coadministration of isoniazid in a patient with AD subject to ahtminium removal by ionspecific chelation. Clin Pharmacol Ther 48: 439, 1990. 35. Flaten T P et al. Desferioxamine for Alzheimer’s Disease. With reply by Andrews D Fetal. Lancet 338 (No 8762): 324f. 1991. 36. Bosman Cl J C G M et al. Alzheimer’s Disease and cellular aging: membrane-related events as clues to primary mechanisms. Gerontology 37: 95. 1991. 37. Rose S P R. How chicks make memories: the cellular cascade from c-fos to dendritic remodelling. Trends in Neurosciences 14: 390, 1991. 38. Reimann J. Alrheimer Erkrankung. AnsPze in der Arrneimitteltherapie. Deutsche Apotheker Zeitung 130: 959, 1990.
I Matid tfypoihma (1992) 38.5-10 0L4mgmnhpuKLtd1992
The Puzzle of Alzheimer’s Disease (AD) H. D. LEHMANN NM Natutwissenschaftliches und Medizinisches lnstitut an der Universit& Tiibingen in Reutlingen, Gustav- Werner-Str 3, D-74 10 Reutlingen, Germany
Abstract-A compromised defensive system of brain cells against aluminium, together with local defects in glucose metabolism, causes AD. Lack of citrate is a driving force and free cis-aconitate or glutamate are potential carriers, which enable the exotoxin to cross lipid membranes. Only a few aluminium ions replace magnesium in key positions. They block the reversibility of phosphorylation reactions, which are important for short term memory: sensitization of the insulin receptor and protein phosphorylations. Due to disturbed phosphorylation of the cytoskeleton, protein synthesis runs out of balance. Efforts to restore the disturbed reactions result in AD specific deposits. Aluminium ions are the common cause for the induction of AD pathogenesis in patients with genetic defects, with mechanical brain lesions or with minor infarcts, as well as with changes in the relation between numbers of neurons and neuron nursing glia cells due to age.
Introduction
have shown AD to be a disease not restricted only to the brain. Together with a discussion of the role of magnesium (6), this may renew interest in the Al hypothesis. There is a possibility that this cation may be the missing link between the different AD research areas, as well as between different types of dementia. The following thoughts are deduced from recent AD literature, as well as from fundamental biochemistry. They have the aim of bringing together the separate AD research areas by a modified aluminium hypothesis and of helping to understand differences or similarities of AD to other dementia.
The study of Alzheimer’s Disease (AD) has become highly specialised. There is a wide variety of different areas of research: the molecular biology of the amyloid precursor protein and how its fragments agglomerate; the deficiencies in brain glucose metabolism; neurotransmitter deficiencies; the research on alterations in neural cytoskeletal proteins, as well as on differences found in fibroblast cell cultures. One hypothesis, discussed controversially some years ago, proposed a role for the exotoxin aluminium (Al) in AD (1, 2), which has neither been proven nor disproven. A recent survey on the association between Alcontaining products and AD stated ‘provocative but inconclusive results’ (3). Some recent publications on transferrin deficienties in AD (4, 5) and on work with non-neural ceils, Date received 14 March 1991 Date accepted 13 November 1991
The protection of brain cells against the exotoxin aluminium
Aluminium is ubiquitous on earth. With food and beverage we all ingest several milligrams of Al every
5
6
day; rather larger amounts are taken in by patients using Al containing antacida. Al is an element with no known biological function and has some well-known toxic properties (7). Why are there no problems with Al normally? How does the organism overcome the permanent challenge of incoming toxic Al? To protect against aluminium toxicity, the organism has developed a system of barriers, to efficiently eliminate and exclude Al. The small AlJ+-cation, with a high charge does not exist in a free form under physiological conditions. It is complexed and has a high affinity to some ligands. Normally, Al is detoxified by forming a complex. There is a distinct hierarchy in the affinity of Al to substances occurring in a biological milieu. Phosphates and carboxylates are the preferred ligands, and the molecules ferritin, transferrin and citrate show the highest affinity and arc used by the organism to detoxify Al. There are three barriers which prevent dietary Alions from intrusion into brain cells: the intestine, the blood-brain barrier and the cell membrane. Most Al ingested leaves the intestines without resorption. Al intruding cells in the intestinal wall is immobilized by ferritin. It leaves the body with the faeces when the cells are peeled off the intestinal wall. Only a small percentage of Al reaches the blood stream. There it is bound to transferrin @KA 14.7) and transported for renal elimination. Under normal conditions the Al-concentration in the blood is, therefore, very low (6ng/ml). Transfer-i-in,the transport vehicle of iron, is able to pass cell membranes by endocytosis at sites where special receptors are available. In the rat, high numbers of these receptors exist in those brain regions which are found sensitive to AD in man (3). Bound to transferrin, Al can pass the blood-brain-barrier by transcytosis. Banks and Kastin (8) report on damages of this barrier due to Al. They found 35 rig/ml Al in the liquor, i.e. an increase of Al behind the bloodbrain-barrier in normal, not demented patients. This reflects the rather low permeability of the cerebrovascular endothelium with its tight junctions. Some Al released from transferrin in the brain may be detoxified by ferritin (9). If there is not sufficient binding capacity of transferrin and ferritin, the next candidate in the affinity hierarchy for complexing with Al is citrate. Its affinity for Al is less than that for transfenin (pKA 14.2). In the system citrate/Al3+/H+ the complex is charged negatively at physiological pH (10). This charge prevents from passing across lipid cell membranes. The citrate concentration in the body fluids is normally high enough to keep Al in a detoxified state, even with high loads, as found in dialysis patients. In
MEDICALHYIWI-HESES
blood, the citrate concentration is about 10-4 mol/l. Citrate is an intermediate of the glucose oxidation process which takes place in the cellular mitochondria (Krebs or citrate cycle). The citrate concentration is higher intracellularly than extracellulary resulting in a gradient over the cell membrane. A citrate leakage out of the cell may be an additional protection for the cell against the intrusion of Al-citrate complexes. The role of Al in AD
Al has long been suspected of playing a role in AD, but the mean concentration levels in brain tissue are similar in AD patients and normal controls (11). Comparable results are also found for serum and liquor. Al-analysis of Alzheimer brains is difficult, either because of the appearance of Al in small focal sites or in very low concentrations. Recent investigations found no accumulation. Most reports in the literature do not state in which complex form the Al is found. Only within affected neurons of Alzheimer patients have high concentrations of Al been found by several laboratories. This result can be understood if both a carrier and a driving force for Al in complexed form transport the ion over the lipid membrane into the intracellular region, i.e. against the normal citrate concentration gradient. This would lead to an intracellular accumulation of Al, when the ligand is metaboliscd within the cell. These conditions are found in several dementia: in AD as well as in Multi-Infarct-Dementia (MID). A local small infarct causes hypoxia and glucose deficiency, which lowers the pH locally. This in turn alters the equilibrium between citrate and its successor products in the Krebs cycle. Citrate is normally dehydrated to cis-aconitate and rehydrated to iso-citrate by aconitase, in a state bound to the enzymes in the mitochondria. There is usually no leakage of cis-aconitate to the cell membrane, as its concentration is too low and the rehydration is rapid at normal pH conditions. In the cytosol, there is an iso-enzyme of isocitrate-dehydrogenase, which splits iso-citrate to CO, and a-ketoglutarate. The Krebs cycle enzyme activity in the mitochondria depends on the cofactor NAD+, the cytosolic iso-enzyme on NADP+. Insufficient NADP+ supply in the cytosol may result in an escape of cis-aconitate under these conditions Cisaconitate escaped from one cell may play a detrimental role as Al-carrier for cells in the neighbourhood. Cis-aconitate probably has a high affinity for Al. This ligand is an appropriate Al-carrier for its transport across the cell membrane, because the complex is not charged and highly lipophilic, i.e. soluble in
7
THE PUZZLE OF AIZHEIMER’S DISEASE
the lipid cell membrane. Cis-aconitate, therefore, is thought to be the pathological Al-carrier in AD and MID. It is generated at low pH values and reaches the cell membrane when the enzyme is defective or its cofactor NADP+ is missing. Guy et al (12) used human neuro-blastoma cells as an in vitro model to show the different carrier activities of citrate, EDTA and malt01 for Al. Rather low values were found for citrate. Another potential carrier for Al may be glutamate, as proposed by Deloncle et al (13). In AD brains, glutamate levels are comparable to normal controls. Glutamate can be used in the Krebs cycle as a substrate for energy gain. Its neurotransmiuer function, together with the normal or even lowered (14) concentrations found in the cerebrospinal fluid of AD patients, make it difficult to understand its detrimental role as an Al-carrier in restricted brain areas. On the other hand, due to its excitotoxic properties, fatal consequences of glutamate on sensitive neurons are known. The energy demand of cells near an infarct-r endothelium cells, neurons or glia cells-will establish a driving force for extracellular metabolizable substrates to enter the cell. Within the cell, the potential Al-carriers cis-aconitate or glutamate are metabolized in the Krebs cycle, if the Al finds another ligand of high affinity. What are the consequences?
In close neighbourhood to the infarct, different cell types become damaged. Within the affinity hierarchy of ligands, Al favours next phosphate groups. Phosphate is present in biological media in rather high amounts, but only certain phosphate assemblies are supposed to have a very high affinity for Al as for instance ATP or inositol 1, 4, S-triphosphate (15). The appropriate geometry within connected groups is given by the dimensions of the cation. AP+ is similar in size to Mg2+. The ion radius for Al3+ with coordination number 6 is 54 pm, the value for Mg2+ with coordination number 4 is 57 pm. (These figures are very similar to the ion radius of Ni2+ in a tetragonal configuration: 55 pm. On Ni-rich soils certain plants detoxify and accumulate this element in high amounts in form of a negatively charged citrate complex). In the absence of citrate, the trivalent Al will complex to phosphate groups which are normally paired with the bivalent cation, Mg2+. This is fatal, because it blocks several vital phosphorylation reactions by inhibiting their reversibility. Reversible phosphorylations need bivalent cations as co-factors having a coordination number of 4, two sites of the maximum coordination
number being left open. This is possible for bivalent Mg but not for trivalent Al with its preferred coordination number of 6. Al with its high affinity, thus, blocks the reversible transfer of phosphate groups. A multitude of fatal consequences results: A.
The glucose supply is stopped as a consequence of the de-activation of the insulin receptor, which is sensitized to glucose by a Mg/Mn-dependent kinase phosphorylation. Reversible tyrosine phosphorylation is important for the function of the insulin receptor. In AD, an altered protein tyrosine phosphorylation is reported (16). Low insulin receptor sensitivity multiplies the effect of a small infarct. It enforces the influx of substances other than glucose to keep the Krebs cycle running, including Al-complexed cis-aconitate and glutamate. Glutamate metabolism is accompanied by transamination reactions. Ammonia is generated endogenously in the brain of Alzheimer patients (17). Recent amino acid investigations indicate selective degeneration of glutamate-sensitive neurons (18, 19). The concentrations of amino-acids occurring in liquor were found to be identical for MID and AD (20).
B.
Reversible phosphorylations of neuronal cytoskeletal proteins will be blocked (21, 22, 23). MAP kinase activated by nerve growth factor is dependant on tyrosine phosphorylation, too. This has been shown on PC 12 cells (24). Differences found in cytoskeletal structures in fibroblasts from Alzheimer patients, compared to controls, are caused by different phosphorylation states (25). In neurons, the phosphorylation of cytoskeletal proteins controls dendritic growth. The reversibility of reactions may be important, especially for the short time memory: information brought via neurotransmitters is transformed by phosphorylation reactions. Irreversibility of this process will result in uncontrolled dendritic growth and also in an overreaction of the proteolytic activities, directed against dendritic growth. This, further, will induce the overreaction of the porteolysis-controlling system, i.e. the synthesis of the amyloid precursor protein by microglia (26). The effect of protein phosphoryiation on amyloid precursor protein generation has been shown (27). Furthermore, Al-induced changes in neuronal chromatin are thought to reduce gene transcription (28, 29). This overreaction of the control system has detrimental consequences. In areas with a local energy deficit, the disintegration of the protease in-
MEDICAL HYPOTHESES
hibition system is incomplete. Ubiquitine, found in AD deposits is a marker substance for incomplete proteolysis. The agglomerating pieces of the precursor protein, together with tau, form the neurofibrils specific for AD. The amyloid is found around neurons with wild dendritic growth, in synapses and within brain blood vessels. Both forms of deposits typical for AD have secondary consequences: Together with the folloing point (C), they induce the psychological manifestations of the disease. C.
Several other phosphorylation reactions are disturbed. The production of neurotransmitters (DOPA; dopamin, serotonin) will be blocked due to the inhibition of the dihydropteridine reductase (30). The inhibition of choline-actyltransferase also results in neurotransmitter deficiencies.
The very slow continuous intrusion of only a few Alions induces a permanent stress situation (31), if the are not detoxified by ferritin (32). Neuronal death due to the neuroexcitatory glutamate is a fatal secondary consequence, as well as the amyloid deposits (33) and the disturbances in neurotransmitter synthesis. According to this modified aluminium hypothesis, three factors create together the conditions for AD on a primary level by blocking the reversibility of the fundamental processes necessary for short term memory. 1.
2.
3.
A malfunction of Al elimination and its detoxifying system: a qualitative or quantitative acquired or inherited defect of transferrin. Removal of Al by appropriate chelating agents, e.g. by deferoxamine (34, 35) seems, therefore, to be a reasonable therapy for AD. Small infarcts result in a local hypoxia, acidosis and a deficiency of glucose. The Krebs cycle supplies become insufficient. Extracellular substrates are used to compensate for the deficiency. Due to a local deficiency of NADPt, cis-aconitate may reach the lipid cell membranes. This may occur in glia cells, which are afflicted by Al, as well as in neurons or endothelial cells and may influence neurons in the neighbourhood.
Can the hypothesis explain special features of AD?
Pathological observations of AD concentrate on the amyloid deposits in the brain. A hypothesis must explain why different cell types are afflicted in focuses
in these areas, although recent observation on cells of non-neuronal origin from Alzheimer patients have shown AD to be a systemic illness, not restricted to the brain (36). There are several possibilities: 1.
2.
3.
In the chicken model, it has been shown (37) that glucose uptake and a following presynaptic protein phosphorylation in distinct brain regions are prerequisites for short term memory. In the evolutionary old part of the brain, the ratio of supplying cells per neuron may be lower than in more recently evolved areas. With age, this ratio is lowered by naturally occurring cell death. It is possible that here a ratio may he reached that is insufficient for normal function. For a slowly proceeding illness, this is, of course, only of interest for cells in the neighbourhood of surviving neurons. This is different to acute brain damage due to neuronal death. Another possibility may lie in the different number of transferrin receptors shown in the rat brain (4).
Another specific problem is that AD is found in two forms: it occurs as a hereditary illness, as well as spontaneously. AD is found as a genetic defect either in combination with Down’s syndrome or occurring in adults (FAD = familial AD). On the other hand, AD occurs in a spontaneous form in the elderly. There is a statistical correlation between the increase in age of our population and the number of AD patients. A high Al load alone, as given in many dialysis patients, is not sufficient to produce AD. Very often AD and MID are combined (38). Mechanical brain lesions are a risk factor for AD (‘dementia pugilistica’). The table shows a potential link between different dementiae. It tries to explain the common appearance of AD in patients with different histories. Although many questions remain unanswered, some controversial features typical for AD can be explained by a hypothesis that proposes that the illness results from an overlap of several causes, linked by Al. The normal protection system of the organism, on the other hand, does not justify a fear of Al nor provide hope for clear results from epidemiologic studies on Al in AD. Conclusion Very small amounts of Al in serum or liquor-amounts not easy to analyze and which do not exceed normal values-are fatal in combination with factors introduced by small infarcts and by age statistics. They block the reversibility of vital phosphorylation reactions if they gain access to the interior of cells. Symptoms of AD can be related directly or indirectly to
THE PUZZU
9
OF AIZHEIMER’S DISEASE
Table Features common
to different dementiae.
Are there Iinks?
some Normal
Dialysis patients
Mulfiinfirctalion
?
n0
Systemically compromised Al-elimination
“0
Locally compromized Krebs Cycle (Glucose and 02-Deficiency)
IlO
no
?
Locally insufficient low relation between glia cells and neurons
no
no
?
High aluminium
no
Resulr
load
“0
In conse-
dialysir patients
?
yes
lowering
Adults with Down Syndmne
with
normal
yes dialysis dementia
“0
MID
the blockage of reversible processes by Al, which replaces Mg-ions at key positions. References 1. Crapper Mclachlan DR. Aluminium and Alzheimer’s disease Neurobiology of Aging 7: 532, 1986. 2. Shore D. Wyatt R 1. Ahtminium and Alzheimer’s Disease. J New Mental Disease 171: 553, 1983. 3. Graves A B et al. The association between aluminium-containing products and Alzheimer’s Disease. J Clin Epidemiol 43: 35, 1990. 4. Pullen R G L et al. Gallitmt-67 as a potential marker for aluminium transpott in rat brain. Implications for Alzheimer’s Disease. J of Neurochemistry 55: 1990. Farrar G et al. Defective gallium-transfertin binding in Alzheimer’s disease and Down Syndrome: possible mechanism for accumulation of aluminium in brain. The Lancet 335: 747, 1990. 5. Connor J R. Iron storage and transport proteins in the brain in aging and Alzheimer’s Disease (abstract). Neurobiology of Aging 11: 77. 1990; Camor J R et al. Cellular distribution of transferrin, ferritin and iron in normal and aged human brains. J Neurosci Res 27:595, 1990. 6. Glick J L. Dcmentias: the role of magnesium deficiency and an hypothesis concerning the patbogenesis of Alzbeimer’s Disease. Medical Hypotheses 31: 211, 1990. 7. Ganrot P 0. Metabolism and possible health effects of aluminium. Enviratmental Health Perspectives 65: 363, 1986. 8. Banks W A, Kastin A J. Aluminium induced neurotoxicity: Alterations in membrane functions at the blood-brain-barrier. Neuroscience and Biobehavioral Reviews 13: 47, 1989. 9. Fleming J. Joshi J G. Ferritin: Isolation of aluminium-ferritin complexes from brain. Biochemistry 84: 7866. 1987. 10. Ghmann L 0. Equilibrium and structural studies of Si IV and Al III in aqueous solution. 17. Stable and metastable complexes in the system H+-A13+-citric acid. Inorg Chem 27: 2565. 1988. 11. Jacobs R W et al. A Reexamination of aluminium in Alzbeimer’s Disease: Analysis by energy dispersive X-ray microprobe and flameless atcmic absorption spectrophotometry. Canad J of Neural Sci 16: 498, 1989.
?
?
In old paiients ?
(inherited)
(inherited)
?
? (due to infarcts)
yes
(yes)
?
? (inherited)
lowering with age
lowering with age
age
yes
guence lo In JVXUl~ head traumata patients
no
I-IO
no
“0
FAD + Down Syndrome
FAD
AD
AD
12. Guy S P et al. Uptake of aluminium by human neuroblastoma cells. Biochem Sot Transactions 18: 392. 1990. 13. Deloncie R et al. Alzheimer’s disease and dementia syndromes consecutive to inbalanced mineral metabolisms subsequent to blood brain barrier alteration (abstract) 1. Symposium at metal ions in biology and medicine, Reims 16th May 1990. in cerebrospinal 14. Basun H et al. Amino acid concentrations fluid and plasma in Alzbeimer’s Disease and healthy control subjects. J Neural Transm 2: 295, 1990. 15. Birchall J D, ChappeII J S. AIuminium. chemical physiology, and Alzheimer Disease. The Lancet 1008, 1988. 16. Shapiro I P et al. Altered protein pbosphotylation in Alxheimer’s Disease. J Neurochem 56: 1154. 1991. 17. Hoyer S et al. Ammonia is endogenously generated in the brain in the presence of presumed and verified dementia of Alzheimer type. Neuroscience Letters 117: 358, 1990. 18. Choi D W. Glutamate neurotoxicity and diseases of the nervous system (Review). Neuron 1: 623, 1988. 19. Lowe S L et al. Ante mortem cerebral amino acid concentrations indicate selective degeration of glutamate-enriched neurons in Alzheimer’s Disease. Nemo Science 38: 571, 1990. 20. Degrell J et al. Amino acid concentrations in cerebrospinal fluid in presenile and senile dementia of Alzheimer type and multi-infarct dementia. Arch Gerontol Geriatr 9: 123. 1989. 21. Blass J P et al. Induction of Alzheimer antigens by an uncoupler of oxidative phosphotylation. Archiv Neural 47: 864, 1990. 22. van Huynh T et al. Reduced protein kinase C immunoreactivity and altered protein phosphorylatiat in Alzheimer’s Disease fib&lasts. Arch Neurol 46: 1195, 1989. 23. Steiner B et al. Phosphorylation of microtubule-associated protein tau: identification of the site for Caz+-calmodulin dependent kinase and relationship with tau phosphorylation in Alzheimer tangles. EMBO J 9: 3539. 1990. 24. Ganez N. Cohen P Dissection of protein kinase cascade by which nerve growth factor activates MAP kinases. Nature 353: 170, 1991. 25. Takeda M et al. Changes in adhesion efficiency and vimetttin distribution of fibroblasts from familial Alzheimer’s Disease patients. Acta Neurologica Scandinavica 82: 238, 1990.
10 26.
27.
28.
29.
30.
31.
32.
MEDICAL HYKYlXE%8
Perlmutter L S et al. Morphologic association between microglia and senile plaque amyloid in Alxheimer’s Disease.. Neuroscience Letters 119: 32. 1990. Buxbaum J D et al. Processing of Alzheimer B/A4 amyloid precursor protein: modulation by agents that regulate protein phosphotylation. Proc NatJ Acad Sci USA. 87: 6003, 1990. Lukiw W J, Crapper McLachlan D R. Chromatin structure and gene expression in Alrheimer’s Disease. Mol Brain Res 7: 227. 1990. Crapper McLachlan D R et al. New evidnece for an active role of aluminium in Alzheimer’s Disease. Canad Journal of Neurol Sci Suppl4, 16: 490, 1989. Ahmann P et al. Serum aluminium levels and erythrocyte dihydropteridine reductase activity in patients on hernodialysis. New England J Med 317: 80, 1987. Papasozomenos S C, Su Y. Altered phosphorylation of Tprotein in heat-shocked rats and patients with Alzheimer’s Disease. Proc Natl Acad Sci USA 88: 4543, 1991. Grundke-Iqbal I et al. Ferritin is a component of the neuritic
(senile) plaque in Alrheimer dementia. Acta Neuropathol81: 105, 1990. 33. Beyreuther K et al. Demenz von Alzheimer Typ. Deutsche Apotheker Zeitung 131: 1414. 1991. 34. Kruck T P A et al. Suppression of deteroxamine mesylate treatment-induced side effects by coadministration of isoniazid in a patient with AD subject to ahtminium removal by ionspecific chelation. Clin Pharmacol Ther 48: 439, 1990. 35. Flaten T P et al. Desferioxamine for Alzheimer’s Disease. With reply by Andrews D Fetal. Lancet 338 (No 8762): 324f. 1991. 36. Bosman Cl J C G M et al. Alzheimer’s Disease and cellular aging: membrane-related events as clues to primary mechanisms. Gerontology 37: 95. 1991. 37. Rose S P R. How chicks make memories: the cellular cascade from c-fos to dendritic remodelling. Trends in Neurosciences 14: 390, 1991. 38. Reimann J. Alrheimer Erkrankung. AnsPze in der Arrneimitteltherapie. Deutsche Apotheker Zeitung 130: 959, 1990.