JOURNAL OF NEUROSCIENCE AND NEUROSURGERY (ISSN:2517-7400)

Whilst controversial, the role of aluminium in Alzheimer’s disease (AD) has not been adequately explained as models of aluminium exposure are not generally available. Renal dialysis (RD) patients are exposed to aluminium through phosphate-binding compounds for controlling hypophosphatemia and because of impaired renal excretion accumulate aluminium in tissues including the brain. These individuals represent a suitable population to study the role of aluminium exposure in inducing AD changes in the brain. We studied the brains of 28 RD individuals, with a history of haemodialysis or peritoneal dialysis, for the presence of AD-type changes and glial pathology, and related these to aluminium exposure and to Apolipoprotein E (APOE) genotype. RD patients had increased Aβ deposition compared to controls but this did not relate to brain aluminium. RD cases carrying the APOEε4 allele were more likely to show Aβ deposition, though cases not carrying the APOEε4 allele also showed Aβ deposition. Neurones immunostaining with AT8 against a phosphorylated epitope of Tau were identified in RD brains though neurofibrillary tangles were not observed. A slight increase in the numbers of microglia was also identified in the brains of RD cases compared to controls. Whilst neuropathological changes suggestive of early AD-type pathology associate with RD, they could not be attributed to aluminium exposure but may be related to the presence of the APOEε4 allele acting in conjunction with other RD pathological processes. Aluminium therefore does not appear to be a major risk factor contributing to AD.

Renal dialysis, Aluminium, Alzheimer’s disease, Astrocytes, Microglia 

AD: Alzheimer’s disease
RD: Renal Dialysis
Aβ: Amyloid beta
NFT: Neurofibrillary tangles
Al: Aluminium
APOE: Apolipoprotein E gene
GFAP: Glial Fibrillary Acidic Protein

Genetic factors and age appear to play a primary role in the development of Alzheimer’s disease (AD) with the Apolipoprotein E gene (APOE) being the main risk factor for late onset sporadic AD [1]. That non-genetic risk factors contribute to the progression of AD cannot be discounted, since numerous studies have shown factors such as head trauma [2], non-steroidal anti-inflammatory drug use [3] and gender [4] have an effect on disease outcome and progression. Though controversial, aluminium (Al) has been suggested by some to play a role in the development of AD, with Al being found to be present in Senile Plaques (SP) and Neurofibrillary Tangles (NFT), the major pathological features of AD [5,6]. The incidence of AD is also suggested to correlate with the levels of Al in the drinking water [7-9]. Biochemical evidence suggests that Al is neurotoxic and can act upon several important pathways involved in neuronal functioning [10]. Whilst this provides circumstantial evidence of Al being involved in AD, no directly relevant animal models to address this issue are available. Patients undergoing dialysis for chronic renal failure may provide a model for studying the effects of long term Al exposure on AD biology. Past use of tap water in haemodialysis procedures led to a fatal neurological disorder, dialysis encephalopathy, which was attributable to the presence of Al in the dialysate fluid [11]. Renal dialysis (RD) patients are still chronically exposed to high levels of blood Al due to the ingestion of large quantities of aluminium containing phosphate binding compounds used to control hyper phosphataemia, and the inability to excrete Al because of their impaired renal function [12]. Whilst overt dementia is not seen in RD patients, cognitive function is impaired, potentially related to renal function and toxaemia [13,14] and CT scan abnormalities are observed which correlate with the amount of Al ingested during treatment [15]. Neuropathologically, such cases show the presence of Amyloid beta (Aβ) immunoreactive SP in about one third of cases, at an age where such changes are not common [16], and also show alterations in Tau protein processing, the major constituent of NFT, which correlate with brain Al levels [17]. Because of the finding of AD-like changes in the brains of RD patients we have quantitatively assessed these changes, along with glial markers, and related these to the levels of Al found in the brains of these cases. Additionally we have also assessed the effect of Apolipoprotein E (APOE) ε4 allele dosage, the major genetic risk factor for sporadic AD, to determine if the presence of this risk factor has an effect on pathology

All studies were approved by the Local Research Ethics Committee. Brains were obtained at post-mortem following informed consent from 28 individuals who had undergone either haemodialysis or continuous ambulatory peritoneal dialysis for chronic renal failure (Table 1). None of the cases had any evidence of dialysis associated encephalopathy/dementia, though one case showed evidence of mild cognitive impairment. Brains from five individuals dying of non-neurological causes and having no history of renal impairment, or of neuropsychiatric disease were used as controls (3 male, 2 female, mean age 63(±8) years). Control cases were not significantly different in age at death from renal dialysis cases (p=0.34, unpaired t-test).Brains were obtained and the tissue processed as described previously [16] (see Supplementary Methods).

Results were analysed using ANOVA and Student t-test where appropriate with p<0.05 being regarded as significant. APOE genotypes were analysed using Chi-squared test with Yates correction for small sample sizes. The study had at 80% power and alpha of 0.05 the ability to detect an effect size of 1.42.

Renal dialysis cases showed increased levels of Aβ deposition compared to the control group, which was significant in the frontal cortex (RD, 0.56% vs control, 0.03%; p=0.033, see Table2). Trends for increased Aβ deposition were apparent in the temporal cortex and hippocampus but these failed to reach significance (Temporal, p=0.12; hippocampus, p=0.16). The deposition of Aβ in RD cases did not appear to correlate with age (r=0.019, p=0.92) or with aluminium concentration (r=0.257, p=0.19), though did show a slight association with APOE genotype, APOE ε4 positive individuals (n=7) showing higher Aβ levels than APOE ε4 negative individuals (n=21) though in no region was this significant. Given the low numbers of control cases used, statistical power to show significant differences may have been a limitation in this study since the study was only powered to detect robust changes (effect size 1.42). Further studies with greater sample sizes for both control and renal dialysis cases will be needed to confirm these changes.

o confirm these changes. Silver staining (Palmgren) failed to reveal the presence of NFT in control or RD cases.No control case showed any evidence of AT8 immunostaining, whilst 17 of the 29 RD cases showed positivity (~58%, p<0.0001), primarily located in the hippocampus though in some cases also in the cortex.This positivity consisted of scattered immunostained pyramidal neurones in the hippocampal CA1 layer and in the parahippocampal gyrus in 8 of the 17 cases.In the remainder of the cases, multiple immunostained neurones and accompanying neurites in the neuropil, were observed throughout the hippocampal formation, primarily in CA1, CA4, and the parahippocampal gyrus. In all cases AT8 immunostaining was of Braak stage 0-1. AT8 immunostaining did not appear to correlate with age or cortical aluminium levels (data not shown). AT8 immunopositive cases showed a slightly lower mean brain aluminium concentration than AT8 negative cases though this was not significant (AT8 positive, 4.5±3.3 mgAl/g tissue; AT8 negative, 6.7±4.2 mgAl/g tissue; p=0.14). Conversely, APOEε4 positive cases showed a slightly, but not significantly, higher AT8 score than APOE ε4 negative cases (ε4 negative, 0.7±0.3; ε4 positive, 2.1±0.8, p=0.13).

There was an increase in the numbers of astrocytes of approximately 20% in frontal and temporal cortex of RD cases compared to controls but this did not reach significance, and in the hippocampus the number of astrocytes was reduced, though again, not significantly so.The numbers of activated astrocytes (increased branching and strong GFAP staining) was increased in the frontal and temporal cortex from RD cases, this increase reaching significance in the temporal cortex (p=0.004, two tailed t-test), though in the hippocampus the number of activated astrocytes remained unchanged. Total microglial numbers were elevated in all areas, though this only reached significance in the temporal cortex (p=0.005, two tailed t-test).Activated microglia were reduced in the frontal cortex and hippocampus and elevated in the temporal cortex but nowhere was this significant. No correlations were observed between astrocyte or microglial numbers and age or brain aluminium in any region studied.

Because of the finding that high aluminium in dialysate can lead to the development of dementia and that, in certain species such as the rabbit, administration of aluminium salts can lead to formation of NFT-like structures in neurones, RD cases have been studied in order to elucidate the role of aluminium in Alzheimer’s disease. The pathology of dialysis encephalopathy/dementia is relatively non-specific with the presence of spongiform change in the upper cortical layers being the only significant finding [18]. Some authors have suggested that argyrophilic inclusions distinct from NFT are present in cortical neurones [19, 20], and that Aβ and senile plaque deposition may also occur [16,21].

In the present study, we have sought to relate quantitative measures of pathology to aluminium exposure and APOE ε4 allele status in a series of RD cases previously shown to have Aβ deposition at an age when this finding is infrequent [16]. In this group of RD cases, overall there was increased deposition of Aβ, though this was not at a density comparable to that found in AD (RD, 0.5%; AD, 4-5%, [22]). This might suggest that RD cases have either increased deposition or reduced clearance of Aβ. The deposition of Aβ was however not related to the burden of Al found in the cortex. There was however a trend towards increased deposition of Aβ in those cases carrying the APOE ε4 allele though this relationship was not distinct since study of individual case details demonstrated that a case with the protective ε2/3 genotype [23] shows neocortical SP at a similar age and brain aluminium content to the more predisposed ε3/4 genotype (e.g. case 3, case 8, supplementary Table 1). In these instances and in ε3/3 cases, APOE genotype alone may not account for the presence of Aβ, indeed, even cases with an AD predisposing APOE genotypes do not always progress to develop AD [24]. It may be that APOE ε4 acting as a predisposing risk factor interacts with the general pathological processes associated with RD to produce Aβ deposition as has been suggested for APOE and head trauma [2,25]. RD patients suffer from a considerable number of metabolic impairments due to their renal failure and there is clear evidence of cognitive impairment [26] and such changes may predispose to accelerated AD changes.

The presence of AT8 immunopositive neurones, indicating hyperphosphorylated Tau may be related to the previous finding of protease-resistant hyperphosphorylated Tau in the brains of this group of RD patients using a biochemical assessment [17,27]. The finding of elevated Tau in the previous study was found to be correlated with aluminium exposure though in this study we were unable to identify any such link. This may be due to the biochemical rather than immunohistochemical approach used, allowing a differentiation between soluble and insoluble forms of Tau. The presence of AT8 immunostaining was however related to the APOE ε4 allele and may indicate a general AD pathological process to be occurring in the brains of RD cases independently of aluminium.

There does not appear to be any major change in the numbers of astrocytes in the brains of RD patients, though in some regions the microglial response appears elevated. This does not however appear to be associated with any inflammatory changes since HLADR expressing microglia appear to be normal or even reduced. One possibility is that the metabolic disturbances because of renal failure which occur prior to and during dialysis [26] affect neuronal or glial integrity which in turn leads to increased microglial numbers. The presence of altered immune function in cases undergoing chronic renal dialysis may however prevent an adequate microglial response from occurring, resulting in increased microglial numbers but a failure to induce HLA-DR expression [28]. Alternatively, the elevation in microglial numbers may relate to the presence of Aβ in some renal dialysis cases, though we were unable to demonstrate any association with Aβ deposition, the APOE ε4 allele and microglial numbers.

Histopathology was performed on the formalin fixed right hemisphere. Previous studies have shown the presence of Aβ deposition as fibrillar aggregates in RD patients but have failed to demonstrate NFT using silver staining methods in RD [16]. Blocks were cut from the superior frontal gyrus (frontal cortex) at the anterior limit of the corpus callosum, from the superior temporal gyrus (temporal cortex) at the level of the amygdala, and from the hippocampus at approximately the level of the lateral geniculate nucleus and embedded in paraffin wax. Sections, 10 µm, were stained with anti-Aβ (clone 6F/3D against 11-17 of Aβ, Dako, UK; 1:100 dilution) to demonstrate Aβ, anti-phosphorylated Tau (clone AT8, Innogenetics, Belgium) to demonstrate NFT and Tau accumulation, anti-GFAP (polyclonal anti-bovine GFAP, Dako, UK; 1:4000 dilution) to demonstrate astrocytes, anti-ferritin (polyclonal anti-human liver ferritin, Dako, UK; 1:1000 dilution) to demonstrate all microglia [29], and with anti-HLA-DR/DP/DQ (clone CR3/43, Dako, UK; 1:100 dilution) to demonstrate MHC class II expression on activated microglia. For each antigen, sections were stained in a single run to minimise potential variation in staining. Sections were dewaxed and taken to water, then for AT8, GFAP, ferritin, and HLA-DR, boiled in a microwave for 10 minutes in 0.1M sodium citrate buffer (pH 6.0). Sections were allowed to cool and then endogenous peroxidise activity quenched by incubation of the sections for 20 minutes in methanol containing 3% hydrogen peroxide. For the demonstration of Aβ, sections were incubated in 98% formic acid for 3 minutes, and the sections subsequently washed extensively in running tap water. Non-specific binding was blocked by incubation of the sections in 1% normal horse serum in Phosphate Buffered Saline (PBS), and the sections then incubated with the primary antibodies at room temperature for 1 hour.Sections were given three five-minute washes in PBS and then incubated with biotinylated secondary antibody (Vectastain Elite, Vector Laboratories, UK) for 30 minutes at room temperature then given three five-minute washes in PBS. Sections were incubated with streptavidin-biotin horseradish peroxidise complex (Vectastain Elite, Vector Laboratories, UK) for 30 minutes at room temperature, given three five-minute washes in PBS, and the antibodies visualised by incubating the sections in 0.25 mg/ml diaminobenzidine tetrahydrochloride in 0.003% hydrogen peroxide in PBS for 10 minutes at room temperature. Sections were then washed extensively in running tap water, given a light haematoxylin counterstain, dehydrated, cleared and mounted.

Aβ load was calculated using imaging densitometry as the percentage grey matter covered by Aβ determined in 3 non-adjacent ribbons of cortex from the pial surface to the white matter, or in three areas of the CA1 pyramidal cell layer of the hippocampus [30]. Because AT8 immunostaining demonstrated neurones, neurites in the neuropil and SP-associated neurites containing NFT/Tau phosphorylated at Ser 202, a semi-quantitative method of assessment was used based on a scale from 0, no staining; 0.5, 1-2 neurones/neurites per cm2 ; 1-5, increasing levels of immunostaining up to levels that are seen in AD cases. Cell counts for total and reactive astrocytes/microglia were made using an eyepiece graticule in 3 nonadjacent ribbons of cortex or CA1 region of the hippocampus.

Brain aluminium levels were determined in frontal cortex as described previously [31] and APOE genotyping was performed using standard methods [32,33].

In summary, we have determined indices of AD neuropathology in renal dialysis cases and attempted to relate them to aluminium exposure. Whilst RD cases show elevated AD-type pathology compared to a control group, and this may be related to the presence of the APOE ε4 allele, no single factor, including aluminium, could account for the elevated AD pathology, although a limitation of the study may be sample size given the multifactorial nature of late onset neurodegenerative disorders. The factors giving rise to AD pathology appear to be complex and several factors may interact to produce the final picture of mildly increased AD pathology.

IKP is supported by funds from Jiwaji University, Gwalior and the Department of Biotechnology, Government of India. We would like to thank Geoffrey Taylor for aluminium determinations. Tissue made available for this study was provided by the Newcastle Brain Tissue Resource which is supported by the Medical Research Council, The Alzheimer’s Research Trust and Alzheimer’s Society through the Brains for Dementia Research Initiative, and through an NIHR Biomedical Research Centre Grant in Ageing and Health. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.

1. Corder EH, Saunders AM, Strittmatter WJ, Schmechel DE, GaskellPC, et al. Gene dose of apolipoprotein E type 4 allele and therisk of Alzheimer’s disease in late onset families. Science. 1993Aug;261(5123):921-923.
   
2. Mayeux R, Ottman R, Maestre G, Ngai C, Tang MX, et al. Synergisticeffects of traumatic head injury and apolipoprotein-epsilon 4in patients with Alzheimer’s disease. Neurology. 1995;45(3 Pt1):555-557.
   
3. BreitnerJC.The role of anti-inflammatory drugs in theprevention and treatment of Alzheimer’s disease. Annu Rev Med.1996;47:401-411.
   
4. Launer LJ, Andersen K, Dewey ME, Letenneur L, OttA, et al. Ratesand risk factors for dementia and Alzheimer’s disease: resultsfrom EURODEM pooled analyses. EURODEM Incidence ResearchGroup and Work Groups. European Studies of Dementia.Neurology. 1999 Jan;52(1):78-84.
   
5. Candy JM, Oakley AE, Klinowski J, Carpenter TA, Perry RH, etal. Aluminosilicates and senile plaque formation in Alzheimer’s disease. Lancet. 1986 Feb;1(8477): 354-357. 
   
6. Perl DP, AR Brody. Alzheimer’s disease: X-ray spectrometricevidence of aluminum accumulation in neurofibrillary tanglebearing neurons. Science. 1980 Apr;208(4441):297-299.
   
7. Flaten TP. Aluminium as a risk factor in Alzheimer’s disease, withe mphasis on drinking water. Brain Res Bull. 2001 May;55(2):187-196.
   
8. Martyn CN, Barker DJ, Osmond C, Harris EC, Edwardson JA,et al. Geographical relation between Alzheimer’s disease and aluminum in drinking water. Lancet. 1989 Jan;1(8629):59-62.
   
9. McLachlan DR, Bergeron C, Smith JE, Boomer D, Rifat SL. et al. Risk for neuropathologically confirmed Alzheimer’s diseaseand residual aluminum in municipal drinking water employingweighted residential histories. Neurology. 1996 Feb;46(2):401-405.
   
10. Savory J, Herman MM, Ghribi O. Mechanisms of aluminum-induced neurodegeneration in animals: Implications for Alzheimer’s disease. J Alzheimers Dis. 2006 Nov;10(2-3):135-144.
    
11. Savory J, Wills MR. Dialysis fluids as a source of aluminumaccumulation. Contrib Nephrol. 1984;38:12-23.
    
12. Alfrey AC.Aluminum metabolism. Kidney Int Suppl. 1986Feb;18:S8-11
    
13. Jackson M , Warrington EK , Roe CJ , Baker LR et al. Cognitivefunction in hemodialysis patients. Clin Nephrol. 1987Jan;27(1):26-30.
    
14. Altmann P, Dhanesha U, Hamon C, Cunningham J, Blair J, et al. Disturbance of cerebral function by aluminium in haemodialysispatients without overt aluminium toxicity. Lancet. 1989Jul;2(8653):7-12.
    
15. Savazzi GM, Cusmano F, Degasperi T. Cerebral atrophy in patientson long-term regular hemodialysis treatment. Clin Nephrol. 1985Feb;23(2):89-95.
16. Candy JM, McArthur FK, Oakley AE, Taylor GA, Chen CP, etal. Aluminium accumulation in relation to senile plaque andneurofibrillary tangle formation in the brains of patients withrenal failure. J Neurol Sci. 1992 Feb;107(2):210-218.
    
17. Wischik CM, McArthur FK, Taylor GA, Edwardson JA, Candy JM,et al. Alzheimer’s-disease-like changes in tau protein processing:association with aluminium accumulation in brains of renaldialysis patients. Lancet. 1994 Apr;343(8904):993-997.
    
18. Winkelman MD, Ricanati ES. Dialysis encephalopathy:neuropathologic aspects. Hum Pathol. 1986 Aug; 17(8):823-833.
    
19. Reusche E. Argyrophilic inclusions distinct from Alzheimerneurofibrillary changes in one case of dialysis-associatedencephalopathy. Acta Neuropathol. 1997 Dec;94(6):612-616.
    
20. Scholtz CL, Swash M, Gray A, Kogeorgos J, Marsh F. Neurofibrillaryneuronal degeneration in dialysis dementia: a feature ofaluminum toxicity. Clin Neuropathol. 1987 May-Jun;6(3): 93-97.
    
21. Brun A, Dictor M, Senile plaques and tangles in dialysis dementia. Acta Pathol Microbiol Scand A. 1981 May;89(3):193-198.
    
22. Singleton AB, Hall R, Ballard CG, Perry RH, Xuereb JH, et al.Pathology of early-onset Alzheimer’s disease cases bearing theThr113-114ins presenilin-1 mutation. Brain. 2000 Dec;123Pt12:2467-2474.
    
23. Corder EH, Saunders AM, Risch NJ, Strittmatter WJ, SchmechelDE, et al. Protective effect of apolipoprotein E type 2 allele for lateonset Alzheimer disease. Nat Genet. 1994 Jun;7(2):180-184.
    
24. Farrer LA, Cupples LA, Haines JL, Hyman B, Kukull WA, et al. Effects of age, sex, and ethnicity on the association betweenapolipoprotein E genotype and Alzheimer disease. A metaanalysis. APOE and Alzheimer Disease Meta Analysis Consortium.JAMA. 1997 Oct;278(16):1349-1356.
    
25. Nicoll JA, Roberts GW, Graham DI. Apolipoprotein E epsilon4 allele is associated with deposition of amyloid beta-proteinfollowing head injury. Nat Med. 1995 Feb;1(2):135-137.
    
26. Tiffin-Richards FE, Costa AS, Holschbach B, Frank RD, Vassiliadou A, et al. The Montreal Cognitive Assessment (MoCA)- a sensitive screening instrument for detecting cognitiveimpairment in chronic hemodialysis patients. PLoS One. 2014Oct;9(10):e106700.
    
27. Harrington CR, Perry RH, Perry EK, Hurt J, McKeith IG, et al.Senile dementia of Lewy body type and Alzheimer type arebiochemically distinct in terms of paired helical filamentsand hyperphosphorylated tau protein. Dementia. 1994 SepOct;5(5):215-228.
    
28. Cohen G, Haag-Weber M, Horl WH. Immune dysfunction inuremia. Kidney Int Suppl. 1997 Nov;62:S79-S82.
    
29. Kaneko Y, Kitamoto T, Tateishi J, Yamaguchi K. Ferritinimmunohistochemistry as a marker for microglia. Acta Neuropathol. 1989;79(2):129-136.
    
30. Hyman BT, West HL, Rebeck GW, Buldyrev SV, Mantegna RN, etal. Quantitative analysis of senile plaques in Alzheimer disease: observation of log-normal size distribution and molecularepidemiology of differences associated with apolipoprotein Egeno type and trisomy 21 (Down syndrome). Proc Natl Acad Sci US A. 1995 Apr;92(8):3586-3590.
    
31. Morris CM, Candy JM, Oakley AE, Taylor GA, Mountfort S, et al.Comparison of the regional distribution of transferrin receptors and aluminium in the forebrain of chronic renal dialysis patients.J Neurol Sci.1989 Dec;94(1-3):295-306.
    
32. Hixson JE, Vernier DT. Restriction isotyping of humanapolipoprotein E by gene amplification and cleavage with HhaI. JLipid Res. 1990 Mar;31(3):545-548.
    
33. Wenham PR, Price WH, Blandell G. Apolipoprotein E genotypingby one-stage PCR. Lancet. 1991 May;337(8750):1158-1159.