Abstract
Background. Senile systemic amyloidosis (SSA) is characterized by deposition of wild‐type transthyretin (TTR)‐based amyloid in parenchymal organs in elderly individuals. Previously, no population‐based studies have been performed on SSA.
Methods. Here we have studied the prevalence and risk factors for SSA in a Finnish autopsied population aged 85 or over, as part of the population‐based Vantaa 85+ Autopsy Study (n = 256). The diagnosis of SSA was based on histological examination of myocardial samples stained with Congo red and anti‐TTR immunohistochemistry. The genotype frequencies of 20 polymorphisms in 9 genes in subjects with and without SSA were compared.
Results. The prevalence of SSA was 25%. SSA was associated with age, myocardial infarctions, the G/G (Val/Val) genotype of the exon 24 polymorphism in the alpha2‐macroglobulin (α2M), and the H2 haplotype of the tau gene (P‐values 0.002, 0.004, 0.042, and 0.016).
Conclusion. This population‐based study shows that SSA is very common in old individuals, affecting one‐quarter of people aged over 85 years. Myocardial infarctions and variation in the genes for α2M and tau may be associated with SSA.
| BMI | = | body mass index (weight divided by height squared) |
| ACE | = | angiotensin‐converting enzyme |
| AD | = | Alzheimer's disease |
| APOE | = | apolipoprotein E |
| BACE2 | = | beta‐amyloid cleaving enzyme2 |
| Df | = | degrees of freedom |
| LPL | = | lipoprotein lipase |
| LRP | = | low‐density lipoprotein receptor‐related protein |
| MI | = | myocardial infarction |
| OR | = | odds ratio |
| PCR | = | polymerase chain reaction |
| SNP | = | single nucleotide polymorphism |
| SSA | = | senile systemic amyloidosis |
| TTR | = | transthyretin |
| α2M | = | alpha2‐macroglobulin |
Introduction
Senile systemic amyloidosis (SSA) is characterized by deposition of wild‐type Citation1 transthyretin (TTR)‐derived amyloid in parenchymal organs, mainly in the heart. Clinical findings of SSA include cardiac failure, conduction disturbances, and arrhythmias Citation2, Citation3, which all are very common among the aged. It has been claimed that SSA significantly contributes to a large proportion of cardiac failures in old individuals and that it is clinically overlooked Citation4. The clinical diagnosis of cardiac amyloidosis rests on findings in electrocardiography, echocardiography, angiography, or technetium scanning Citation5, Citation6, but the definite diagnosis of amyloidosis requires microscopic examination of a biopsy sample, preferentially from the myocardium or alternatively from another tissue, e.g. the subcutaneous fat or rectum Citation7, Citation8. This makes the diagnosis difficult especially in aged individuals with increased risks in invasive operations. Thus, the diagnosis of SSA is mostly set by chance at the autopsy Citation2, Citation4, Citation7. The actual prevalence of SSA has remained unknown due to lack of population‐based studies on this condition.
Except for the effect of age Citation9 little is known of the genetic or other risk factors for SSA. In a previous hospital‐based autopsy study, a weak tendency for an association between the myocardial infarctions (MIs) and SSA has been proposed Citation9. In addition, it has been suggested that males would be more prone to develop SSA than females Citation3, Citation4 because of the overrepresentation of male patients in hospital‐based studies Citation4, Citation7.
In this study, we have analyzed the prevalence and risk factors of SSA in a population‐based Finnish autopsy sample (the Vantaa 85+ Study). This sample is well characterized, and has been successfully used to identify risk factors for other age‐related disorders, including Alzheimer's disease (AD) and cerebral amyloid angiopathy Citation10–14.
Key messages
The prevalence of senile systemic amyloidosis (SSA) in the autopsied elderly population is high, 25%.
Age, myocardial infarctions, and genetic variation in the genes for alpha2‐macroglobulin and tau associate with SSA, and may influence its pathogenesis.
Methods
Study population and setting
This study is part of the population‐based, prospective Vantaa 85+ Study, which includes all the individuals (n = 601) living in the city of Vantaa, southern Finland, on April 1, 1991, born earlier than April 1, 1906. The study participants were interviewed and clinically examined in the baseline study in 1991 as well as in the follow‐ups in 1994, 1996, and 1999 Citation15. The present study is based on the results of 256 consented postmortem examinations. The age range for the study subjects was 85.1–105.6 years (mean 92.5 years), and 84% (214/256) were females, 16% (42/256) males. There were no significant differences in age or gender between this autopsied subpopulation and the whole study population.
Histological examination
At autopsy one standard specimen of the interventricular septum of the heart was taken, with two to five additional samples if the macroscopic findings were abnormal. The specimens were fixed in phosphate‐buffered 4% formaldehyde solution. SSA was diagnosed by detecting amyloid on the Congo red‐stained specimens by typical red–apple green birefringence of amyloid in polarized light. Based on the Congo red staining, the amount of amyloid in the myocardium was semiquantitatively graded: 0 = no amyloid; 1 = small amounts of amyloid in the vascular walls or between the heart muscle cells; 2 = clearly detectable areas of amyloid in several visual fields, including vascular deposits; and 3 = large amounts of amyloid.
Immunohistochemistry
Characterization of the main amyloid protein was performed immunohistochemically on all Congo red‐positive samples. The sections were stained by using Avidin‐Biotin Complex Method using Vectastain® ABC kit and a monoclonal antibody recognizing normal TTR (Dako, Glostrup, Denmark, in dilution 1:1000).
The health‐associated risk factors
The information on diabetes, hypertension, smoking, and body mass index (BMI; body mass index = weight divided by height squared) was included in the Vantaa 85+ Study protocol, and was obtained in the clinical baseline study and the follow‐ups Citation15. In the study protocol, subjects were classified as diabetics if they used blood glucose‐lowering medication or insulin, and as hypertensive if they used any blood pressure‐lowering medication. The BMI values were classified into five categories: 1⩾32; 2 = 28–31; 3 = 19–27; 4 = 15–18; 5⩽14. The BMI category was estimated, when exact measurement was not possible. Weight of the heart was measured at the autopsy. The diagnosis of MI was made at the autopsy with naked eye and subsequently confirmed with histological examination. Both acute infarctions and old scars were noted.
Genetic analyses
In order to identify possible genetic risk factors for SSA, we studied the same candidate gene polymorphisms, which we had previously assessed in studies concerning other age‐related disorders, e.g. AD and cerebral or myocardial infarction. There were 19 such polymorphisms in 8 different genes: alpha2‐macroglobulin (α2M), angiotensin‐converting enzyme (ACE), apolipoprotein E (APOE), beta‐amyloid cleaving enzyme2 (BACE2), lipoprotein lipase (LPL), low‐density lipoprotein receptor‐related protein (LRP), prion protein, and tau. The polymorphisms were analyzed by polymerase chain reaction (PCR)‐based methods Citation11–13, Citation15–19 as described elsewhere. The genotype frequencies of these polymorphisms in subjects with and without SSA were compared.
In addition to examining a possible association between these polymorphisms and SSA, the coding sequence of the TTR gene of six individuals with severe (grade 3) SSA was sequenced to test whether or not the coding sequence of the TTR gene itself was abnormal in these subjects. The sequencing was performed by amplifying the exons of the TTR gene by standard PCR methods (primers and PCR conditions available on request). PCR products were purified using enzymatic ExoSAP treatment (USB Corporation, Cleveland, OH, USA) and sequenced using BigAnalyzed (Applied Biosystems, Foster City, CA, USA). The analysis of the sequences was carried out with Sequencer 4.5 Software (Gene Codes Corporation, Ann Arbor, MI, USA). Genotyping of the mutation in exon 2 was done in all samples by using the same PCR primers and conditions which were used for sequencing of the exon 2. The PCR product was digested with restriction enzyme Msp I (New England Biolabs, Ipswich, MA, USA).
Statistical analyses
The statistical analyses were performed by using the SPSS for Windows version 12.0.1 software. The differences between the groups were analyzed with chi‐square or exact test for linear trend for categorical variables, and with logistic regression analysis for continuous variables. Associations between the genotypes and severity of SSA (grades 1–3) were studied by using chi‐square, exact test for linear trend, or Fischer's exact test. Logistic regression analysis was used to control for possible confounding effects of age and MI on the association between the genotypes and SSA. P‐value <0.05 was considered significant.
Approval for the study
The ethical committee of the City of Vantaa has approved the study. A written consent from the nearest relative for the autopsy was obtained.
Results
Prevalence of SSA
Characteristics of the study population and the number of individuals with and without SSA are shown in . Amyloid deposition in one or more of the Congo red‐stained myocardial specimens was detected in 25% (63/256) of the study population. In all cases the amyloid deposits showed a positive reaction in the anti‐TTR immunohistochemistry. SSA was mild in 78% (49/63), moderate in 11% (7/63), and severe in 11% (7/63). The proportion of individuals with SSA clearly increased along with the age at death of the subjects. SSA was detected in 17% (11/65) of the individuals who died at the age of 85–89.9 years, in 23% (29/127) of those who died at the age of 90–94.9 years, in 32% (18/56) of those who died at the age of 95–99.9 years, and in 63% (5/8) of those reaching the age of at least 100 years ().
Table Ia. The study subjects stratified by age at death, gender, and grade of senile systemic amyloidosis.
The health‐associated risk factors
SSA was significantly associated with the age at death (OR 1.13; 95% CI 1.05–1.22; P = 0.002) and with the heart weight at the autopsy (OR 1.004; 95% CI 1.001–1.007; P = 0.015) (). Although the frequency of males with SSA did not differ significantly from that of females, the severity of SSA was significantly associated with male gender (P = 0.022). Particularly, the severe (grade 3) SSA occurred in 10% of the males but in only 1% of the females. Low BMI and MIs showed a nonsignificant tendency towards association with SSA (P = 0.053 and P = 0.076, respectively). SSA was not associated with diabetes, hypertension, or smoking ().
Table Ib. Demographic and health‐related characteristics of individuals with and without senile systemic amyloidosis.
Candidate gene analysis
shows the results of the association analyses comparing the genotype distributions of subjects with and without SSA. SSA associated significantly with the exon 24 polymorphism of the α2M gene (P = 0.042) and with the exon 9 deletion/insertion polymorphism of the tau gene (P = 0.016). We next studied whether the severity (grade) of SSA and the distribution of α2M and tau genotypes were associated (). The exon 24 polymorphism of α2M as well as the tau H1/H2 haplotype showed significant association with the severity of SSA (P = 0.012 and P = 0.003, respectively).
Table II. The results of association analyses comparing the genotype distributions in subjects with and without senile systemic amyloidosis.
Table III. Comparison of genotypes of the alpha2‐macroglobulin exon 24 and tau H1/H2 polymorphisms in subjects classified according to the severity (grade) of senile systemic amyloidosis.
Multivariate analysis of age, MI, and the genetic variants
We next studied the association between SSA and α2M and tau by controlling for the possible confounding effects of age and MIs in logistic regression analysis (). The P‐values for α2M and tau remained significant when age at death and MIs were included in the analysis, indicating that the effects of tau and α2M genotypes were independent of the other variables. In addition, adjusting for other variables revealed that MIs were significantly associated with SSA.
Table IV. Results of a multivariate analysis assessing the effects of age, myocardial infarctions, and the genetic variants of alpha2‐macroglobulin and tau on senile systemic amyloidosis.
Sequencing of the coding region of the TTR gene
The coding regions of the TTR gene in six individuals with severe (grade 3) SSA were sequenced. The analysis showed predicted wild‐type sequence in five cases, but one individual was heterozygous (A/G genotype) for a single nucleotide polymorphism (SNP) in exon 2 on the TTR gene. This SNP is found in the database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD = search&DB = snp) as rs1800458, and it is predicted to result in a nonsynonymous amino acid change Gly26Ser, previously known as Gly6Ser Citation20.
Frequency of the Gly26Ser mutation of the TTR gene
The nonsynonymous SNP rs1800458 was subsequently analyzed in all samples of the study. We found 2 homozygous (Ser/Ser) and 43 heterozygous (Gly/Ser) individuals, while 151 subjects were homozygous with the wild‐type allele. The allele frequency in the population was 0.12. There was no difference in the frequency of this variant when subjects with and without SSA were compared.
Discussion
Our results showed that SSA is a very common condition, affecting one‐quarter of the very aged individuals. Thus, our population‐based study is in accordance with the previous hospital‐based studies in which SSA was detected in 25% of autopsied subjects 80 years of age or over Citation9, Citation21. In addition, we showed a strong association between SSA and age at death and provided evidence for increasing occurrence of SSA along with increasing age at death. Interestingly, the severity of SSA was associated with male gender, although the overall frequency of SSA did not differ significantly between genders. It is possible that the hospital‐based studies, which suggested that SSA is more common in males than in females Citation3, Citation4, Citation7, have suffered from a selection bias because the more severe myocardial amyloidosis in males is more likely to be detected than the mild disease in females.
We also found a trend for an association between MIs and SSA, analogous with a previous study Citation9, and this association reached significance after adjusting for age and the genetic risk factors. To the best of our knowledge, the mechanism between SSA and MIs has not been studied. One possibility is that the MIs are a secondary phenomenon due to the obliteration of cardiac blood vessels by amyloid; a similar link between vascular amyloid and microinfarctions has been noted in cerebral amyloid angiopathy Citation22. Increased heart weight at the autopsy, a common observation in various myocardial diseases, was also associated with SSA. Thus, our findings indicate that SSA often combines with other myocardial pathology, and this may be one of the reasons why SSA is clinically so infrequently diagnosed.
In this study, the occurrence of SSA was determined by examining cardiac samples only, because the heart is the organ most commonly affected by SSA Citation2. As SSA is also seen in other tissues, preferably in the lungs, we also examined the presence of pulmonary SSA in pulmonary sections, which were available in 164 out of the total 256 individuals. Among these 164 people, there were only 2 subjects (1%) in whom amyloid was detected in the pulmonary blood vessels but not in the heart. Thus, determining the occurrence of SSA by examining cardiac samples seems to be a reliable method.
In a previous genetic study on SSA, the coding sequence of TTR was sequenced in two Swedish patients, but no mutations were found Citation1. We now extended this study by performing a similar analysis on six individuals with severe (grade 3) SSA in the study population. Five of them had the predicted wild‐type sequence, but one individual carried a minor allele of the SNP rs1800458. The frequency of this variant was 0.12 in our population, and the genotype frequencies did not differ when comparing subjects with and without SSA. Thus, our data suggest that variation in the TTR gene does not play a major role in the genetic predisposition to sporadic SSA in the Finnish population.
The candidate gene screen showed a significant association between SSA and polymorphisms for the α2M and tau genes in our study population. With the exception of α2M, none of the genetic polymorphisms that in our previous studies were associated with AD, associated with SSA. These findings suggest that the pathogenic pathways leading to SSA and AD may differ from each other. Moreover, as shown in our previous study Citation23, the pathologies in SSA and AD were not associated with each other in subjects aged 95 or over.
α2M is a pan‐proteinase inhibitor, found in the senile plaques of AD Citation24, in the tissue deposits of AA (inflammation‐associated, formerly called secondary) amyloidosis Citation25, and in the affected tissues in the beta2‐microglobulin (dialysis)‐associated amyloidosis Citation26. Although the exact mechanism how α2M affects the accumulation of amyloid is not known, it has been suggested that it may facilitate conformational changes in the amyloid protein Citation27. Genetic associations between AD and an intronic SNP upstream from exon 18 and the Ile1000Val SNP in exon 24 of the α2M gene have been reported Citation11, Citation28–30. Interestingly, the genotype in exon 24 of α2M, which in this study associated with SSA, was G/G (Val/Val), whereas the genotype that associated with AD in two separate Finnish data sets was A/A (Ile/Ile) Citation11, Citation30. Thus, the α2M variant, which increases the risk for AD, seems to have an opposite effect on SSA. This is consistent with the lack of association between the SSA and AD phenotypes, reported previously Citation23.
Tau is a microtubule‐associated protein, found in its phosphorylated form in the nervous tissue in many neurodegenerative diseases, including AD Citation31, as well as in amyloid deposits in sporadic inclusion body myositis Citation32. Tau has a well known microtubule‐stabilizing function, and it has been shown to interact with actin Citation33. However, its exact function in amyloid formation remains unknown. Tau gene on chromosome 17 is covered by two extended haplotypes H1 and H2, of which H1 is more common Citation34. Variation in the tau gene has been associated with several neurodegenerative diseases Citation19, Citation31, but before this study the tau haplotypes or mutations have not been associated with any nonneurological disorder. Interestingly, the tau H1/H2 polymorphism is an inversion polymorphism, and the inversed region contains several other genes as well Citation35. Thus, it remains possible that the association found between the tau haplotype and SSA reflects an effect of another gene in that genetic region.
Owing to the difficulty of the clinical diagnosis of SSA, an autopsy study is in practice the only method to define its prevalence. The high autopsy rate (approximately 50%), compared with other similar studies Citation36, guarantees the reliability of the diagnostics in this study. Furthermore, our prospective and population‐based study setting rules out the selection bias often involved in hospital‐based or case‐control studies. Although population‐based studies have been underemphasized, it has been recognized that they may provide unique insights in age‐associated diseases by defining the prevalence of pathology and genetics at a population level Citation36. Finally, as the Finnish population is based on a small number of founders, the number of genetic risk factors of diseases may be smaller than in other populations, making our study powerful.
AD, the best known of the age‐related amyloid diseases, has been associated with several genetic polymorphisms, but the ϵ4 allele of APOE is the only widely accepted risk factor for AD. Thus far, apart from age, no genetic or other risk factors have been established for SSA, and further studies are needed to confirm the involvement of α2M and tau in its pathogenesis. Clinically, SSA has been an underdiagnosed condition, but its high prevalence and the association with the MIs proven here necessitate prospective clinical studies to define its true clinical implication among the elderly.
Acknowledgements
We thank Pirjo Tuomi and Tuija Järvinen for skillful technical assistance. The study was supported by the Red Feather Project of the Lions, by Päivikki and Sakari Sohlberg Foundation, Finnish Cultural Foundation, Helsinki University Central Hospital, and by the Finnish Medical Foundation.
References
- Gustavsson A., Jahr H., Tobiassen R., Jacobson D. R., Sletten K., Westermark P. Amyloid fibril composition and transthyretin gene structure in senile systemic amyloidosis. Lab Invest 1995; 73: 703–8
- Pitkanen P., Westermark P., Cornwell G. G 3rd. Senile systemic amyloidosis. Am J Pathol 1984; 117: 391–9
- Johansson B., Westermark P. Senile systemic amyloidosis: a clinico‐pathological study of twelve patients with massive amyloid infiltration. Int J Cardiol 1991; 32: 83–92
- Ng B., Connors L. H., Davidoff R., Skinner M., Falk R. H. Senile systemic amyloidosis presenting with heart failure: a comparison with light chain‐associated amyloidosis. Arch Intern Med 2005; 165: 1425–9
- Stone M. J. Amyloidosis: a final common pathway for protein deposition in tissues. Blood 1990; 75: 531–45
- O'Hara C. J., Falk R. H. The diagnosis and typing of cardiac amyloidosis. Amyloid 2003; 10: 127–9
- Kyle R. A., Spittell P. C., Gertz M. A., Li C. Y., Edwards W. D., Olson L. J., et al. The premortem recognition of systemic senile amyloidosis with cardiac involvement. Am J Med 1996; 101: 395–400
- Hughes S. E., McKenna W. J. New insights into the pathology of inherited cardiomyopathy. Heart 2005; 91: 257–64
- Cornwell G. G 3rd., Murdoch W. L., Kyle R. A., Westermark P., Pitkanen P. Frequency and distribution of senile cardiovascular amyloid. A clinicopathologic correlation. Am J Med 1983; 75: 618–23
- Polvikoski T., Sulkava R., Haltia M., Kainulainen K., Vuorio A., Verkkoniemi A., et al. Apolipoprotein E, dementia, and cortical deposition of beta‐amyloid protein. N Engl J Med 1995; 333: 1242–7
- Myllykangas L., Polvikoski T., Sulkava R., Verkkoniemi A., Crook R., Tienari P. J., et al. Genetic association of alpha2‐macroglobulin with Alzheimer's disease in a Finnish elderly population. Ann Neurol 1999; 46: 382–90
- Myllykangas L., Polvikoski T., Sulkava R., Verkkoniemi A., Tienari P., Niinisto L., et al. Cardiovascular risk factors and Alzheimer's disease: a genetic association study in a population aged 85 or over. Neurosci Lett 2000; 292: 195–8
- Myllykangas L., Wavrant‐De Vrieze F., Polvikoski T., Notkola I. L., Sulkava R., Niinisto L., et al. Chromosome 21 BACE2 haplotype associates with Alzheimer's disease: a two‐stage study. J Neurol Sci 2005; 236: 17–24
- Tanskanen M., Lindsberg P. J., Tienari P. J., Polvikoski T., Sulkava R., Verkkoniemi A., et al. Cerebral amyloid angiopathy in a 95+ cohort: complement activation and ApoE genotype. Neuropathol Appl Neurobiol 2005; 31: 589–99
- Myllykangas L., Polvikoski T., Sulkava R., Notkola I. L., Rastas S., Verkkoniemi A., et al. Association of lipoprotein lipase Ser447Ter polymorphism with brain infarction: a population‐based neuropathological study. Ann Med 2001; 33: 486–92
- Nurmi M. H., Bishop M., Strain L., Brett F., McGuigan C., Hutchison M., et al. The normal population distribution of PRNP codon 129 polymorphism. Acta Neurol Scand 2003; 108: 374–8
- Baker M., Graff‐Radford D., Wavrant DeVrieze F., Graff‐Radford N., Petersen R. C., Kokmen E., et al. No association between TAU haplotype and Alzheimer's disease in population or clinic based series or in familial disease. Neurosci Lett 2000; 285: 147–9
- Myllykangas L., Polvikoski T., Reunanen K., Wavrant‐De Vrieze F., Ellis C., Hernandez D., et al. ApoE epsilon3‐haplotype modulates Alzheimer beta‐amyloid deposition in the brain. Am J Med Genet 2002; 114: 288–91
- Myers A. J., Kaleem M., Marlowe L., Pittman A. M., Lees A. J., Fung H. C., et al. The H1c haplotype at the MAPT locus is associated with Alzheimer's disease. Hum Mol Genet 2005; 14: 2399–404
- Jacobson D. R., Alves I. L., Saraiva M. J., Thibodeau S. N., Buxbaum J. N. Transthyretin Ser 6 gene frequency in individuals without amyloidosis. Hum Genet 1995; 95: 308–12
- Westermark P., Bergstrom J., Solomon A., Murphy C., Sletten K. Transthyretin‐derived senile systemic amyloidosis: clinicopathologic and structural considerations. Amyloid 2003; 10 Suppl 1: 48–54
- Vinters H., Wang Z., Secor D. Brain parenchymal and microvascular amyloid in Alzheimer's disease. Brain Pathol 1996; 6: 179–95
- Tanskanen M., Kiuru‐Enari S., Tienari P., Polvikoski T., Verkkoniemi A., Rastas S., et al. Senile systemic amyloidosis, cerebral amyloid angiopathy, and dementia in a very old Finnish population. Amyloid 2006; 13: 164–9
- Rebeck G. W., Harr S. D., Strickland D. K., Hyman B. T. Multiple, diverse senile plaque‐associated proteins are ligands of an apolipoprotein E receptor, the alpha 2‐macroglobulin receptor/low‐density‐lipoprotein receptor‐related protein. Ann Neurol 1995; 37: 211–7
- Muller D., Roessner A., Rocken C. Distribution pattern of matrix metalloproteinases 1, 2, 3, and 9, tissue inhibitors of matrix metalloproteinases 1 and 2, and alpha 2‐macroglobulin in cases of generalized AA‐ and AL amyloidosis. Virchows Arch 2000; 437: 521–7
- Campistol J. M., Shirahama T., Abraham C. R., Rodgers O. G., Sole M., Cohen A. S., et al. Demonstration of plasma proteinase inhibitors in beta 2‐microglobulin amyloid deposits. Kidney Int 1992; 42: 915–23
- Adler V., Davidowitz E., Tamburi P., Rojas P., Grossman A. α2‐Macroglobulin is a potential facilitator of prion protein transformation. Amyloid 2007; 14: 1–10
- Blacker D., Wilcox M. A., Laird N. M., Rodes L., Horvath S. M., Go R. C., et al. Alpha‐2 macroglobulin is genetically associated with Alzheimer disease. Nat Genet 1998; 19: 357–60
- Liao A., Nitsch R. M., Greenberg S. M., Finckh U., Blacker D., Albert M., et al. Genetic association of an alpha2‐macroglobulin (Val1000lle) polymorphism and Alzheimer's disease. Hum Mol Genet 1998; 7: 1953–6
- Pirskanen M., Alafuzoff I., Hiltunen M., Mannermaa A., Helisalmi S., Lehtovirta M., et al. An association between a subset of Finnish late‐onset Alzheimer's disease and alpha2‐macroglobulin. Neurogenetics 2001; 3: 171–2
- Goedert M., Jakes R. Mutations causing neurodegenerative tauopathies. Biochim Biophys Acta 2005; 1739: 240–50
- Askanas V., Engel W. K., Bilak M., Alvarez R. B., Selkoe D. J. Twisted tubulofilaments of inclusion body myositis muscle resemble paired helical filaments of Alzheimer brain and contain hyperphosphorylated tau. Am J Pathol 1994; 144: 177–87
- Ballatore C., Lee V. M., Trojanowski J. Tau‐mediated neurodegeneration in Alzheimer's disease and related disorders. Nat Rev Neurosci 2007; 8: 663–72
- Baker M., Litvan I., Houlden H., Adamson J., Dickson D., Perez‐Tur J., et al. Association of an extended haplotype in the tau gene with progressive supranuclear palsy. Hum Mol Genet 1999; 8: 711–5
- Stefansson H., Helgason A., Thorleifsson G., Steinthorsdottir V., Masson G., Barnard J., et al. A common inversion under selection in Europeans. Nat Genet 2005; 37: 129–37
- Zaccai J., Ince P., Brayne C. Population‐based neuropathological studies of dementia: design, methods and areas of investigation—a systematic review. BMC Neurol 2006; 6: 2