Content uploaded by Christine Van Broeckhoven
Author content
All content in this area was uploaded by Christine Van Broeckhoven on Mar 04, 2018
Content may be subject to copyright.
A preview of the PDF is not available
Genetic linkage studies suggest that Alzheimer's disease is not a single homogeneous entity
Abstract and Figures
Alzheimer's disease, a fatal neurodegenerative disorder of unknown aetiology, is usually considered to be a single disorder because of the general uniformity of the disease phenotype. Two recent genetic linkage studies revealed co-segregation of familial Alzheimer disease with the D21S1/S11 and D21S16 loci on chromosome 21. But two other studies, one of predominantly multiplex kindreds with a late age-of-onset, the other of a cadre of kindreds with a unique Volga German ethnic origin, found absence of linkage at least to D21S1/S11. So far it has not been possible to discern whether these conflicting reports reflect aetiological heterogeneity, differences in methods of pedigree selection, effects of confounding variables in the analysis (for example, diagnostic errors, assortative matings), or true non-replication. To resolve this issue, we have now examined the inheritance of five polymorphic DNA markers from the proximal long arm of chromosome 21 in a large unselected series of pedigrees with familial Alzheimer's disease. Our data suggest that Alzheimer's disease is not a single entity, but rather results from genetic defects on chromosome 21 and from other genetic or nongenetic factors.
Figures - uploaded by Christine Van Broeckhoven
Author content
All figure content in this area was uploaded by Christine Van Broeckhoven
Content may be subject to copyright.
... Examples of genetic heterogeneity in human diseases include polycystic kidney disease (Reeders et al., 1987), early-on set Alzheimer's disease (St. George-Hyslop et al, 1990;, maturity-onset diabetes of the young (Froguel et al., 1992), hereditary nonpolyposis colon cancer (Fishel et al., 1994), and ataxia telangiectasia (Sobel et al., 1992), etc. Medical geneticists are normally unable to know whether two patients suffer from the same disease for different genetic reasons until the responsible mutations are recognized. A chromosome region (locus) may cosegregate with a disease in some families but not in others reflecting heterogeneity of linkage. ...
... Alzheimer's disease is also genetically heterogeneous (St George-Hyslop, 1990; Van Broeckhoven et al, 1992). This is an example to show that over zealous exclusion of strong candidate genes is not wise for a genetically heterogeneous diseases. ...
Schizophrenia is a severe disabling neuropsychiatric syndrome affecting 0.85% of most human populations. Family, twin, and adoption studies in several ethnic populations have consistently implicated genetic factors as having an important pathogenic role in the disease. However, researchers have only recently begun to detect consistent evidence of linkage between schizophrenia and genetic markers, but no candidate genes have yet been implicated. It has been hypothesized that glutamate receptor function is important in either the aetiology or treatment of schizophrenia and therefore the glutamate receptor family of genes are potential susceptibility loci for schizophrenia. To test this hypothesis, twenty-three English and Icelandic schizophrenia families containing multiple cases of schizophrenia were genotyped with currently available microsatellite polymorphisms localized at the GluR5, GluR6 glutamate receptors and SLC1A5 glutamate/aspartate transporter loci. Lod scores, model-free linkage analysis, and extended relative pair analysis methods were used to test for linkage. No evidence of close linkage between schizophrenia and any of these loci was found. In addition, in order to understand how specific glutamate receptor genes are involved in the treatment of schizophrenia, a multiprobe oligonucleotide solution hybridization (MOSH) technique was used to examine the regulation of gene expression of the ionotropic glutamate receptor subunit genes. Four regions of the left rat brain following treatment with the optical isomers of flupenthixol at a dose of 0.2 mg/kg/day over a period of 1, 2, 4, 8, 12, 24 weeks were studied. A previous controlled trial showed that cis-flupenthixol had a significantly superior ability to ameliorate the positive symptoms of schizophrenia compared to its transisomer. Protein levels of the NMD A receptor subunit NR1, GluR2/3 and GluR6/7 glutamate receptors in the right brain were also examined by Western blotting technique with specific antibodies for these receptor subunits. In summary, the gene expression of specific NMD A receptor subunits in several regions of the left rat brain was altered by treatment with either the cis- or transisomer of the antipsychotic drug flupenthixol. NRl mRNA was significantly decreased throughout the 24 weeks treatment with trans-flupenthixol and after long-term (12 or 24 week) treatment with cis-flupenthixol in the frontal and subcortical areas. NR2B and NR2C mRNA expression demonstrated a dynamic pattern of change in different brain regions following treatment with flupenthixol whilst NR2A and NR2D gene expression was relatively unaffected except in the subcortical region. The gene expression of AMPA (GluRl-4) and kainate (GluR5-7, KA1, KA2) types of glutamate receptor subunits was unaffected following 4 and 24 weeks of treatment with either trans- or cis-flupenthixol. It would be difficult to make inferences about the pathophysiology of schizophrenia or any other psychiatric disorders solely based on drug mechanisms. However, these results indicate that adaptations in glutamate receptors may represent an important and novel mechanism through which neuroleptics exert some of their effects on brain function.
... A small percentage of the population (less than 1%) have an inherited autosomal dominant form known as familial AD (FAD). Genetic studies have established that AD is aetiologically heterogeneous (St George- Hyslop et al., 1990), and that this may contribute to the variation found in clinicopathological findings. 31 ...
... A point mutation caused a single amino acid substitution (glutamine to glutamic acid) at position 2 2 of Ap that was likely to be the reason for the deposition of Ap in the cerebral wall vessels of these patients, followed by cerebral haemorrhage and premature death. Linkage of chromosome 21 with late-onset AD was not observed (Pericak-Vance et al., 1988;St George-Hyslop et al., 1990), but linkage with early-onset AD was seen in some families (Goate et al., 1989). Following the clues that exon 17 of APP played a part in HCHWA-D another group looked for and found a point mutation in this exon in AD . ...
Quantitative human mRNA data are derived from post-mortem or biopsied tissue. Confounding factors, RNA degradation, poor replication and a large variance are often cited, however, as objections to the data's reliability. At issue is whether post-mortem mRNA represents an ordered system and to what degree non-specific factors contribute to the measurements. I developed statistical methods and validated them by measuring 25 mRNA transcripts in an animal model of ischaemia. In the process I discovered novel increases for 3 genes in rats with ischaemic damage: leukaemia inhibitory factor, nestin and galanin mRNA. Additionally, I discovered that reference genes known as "housekeepers"' do not always act as steady-state controls and that the precise value of a test gene response varies according to the baseline choice of reference gene. Once optimised, I applied the analytical methods to human post-mortem brains. I used TaqMan™ real-time RT-PCR to measure 13 mRNAs in 513 cortical samples taken from 90 Alzheimer's disease and 81 control brains. Despite a high variance and confounding factors such as brain pH, I found strong geometric relations between the mRNA transcripts up to and beyond 100 hours autopsy delay. Where a postmortem brain had a high/low level of one mRNA, the same brain invariably had a high/low level of other mRNAs; correlated order is present and provides a means of isolating any mRNA change due specifically to disease. I measured mRNA levels of β-Secretase (BACE), GSK 3 and the isoforms of APP/APRP in the AD and control brains. After adjustment for age of death, brain pH, and gender, there was no change in the mRNA levels for either BACE or GSK 3α mRNA (p = 0.354 and p = 0.054 respectively). There was a change, however, in the ratio of KPI+ to KPI- mRNA isoforms of APP/APRP. Three separate probes, designed only to recognise KPI+ mRNA, each gave increases of between 28 and 50% in AD brains relative to controls (p = 0.002). There was no change in the mRNA levels of KPI-(APP 695) (p = 0.898). Therefore, whilst I KPI- mRNA levels remained level between AD and control brains, the KPI+ species were seen to increase specifically in the AD brains.
... Early genetic studies showed that only a minority of cases of FAD could be explained by mutations at chromosome 21 as the majority of cases did not segregate with an APP mutation [243][244][245] . In fact, linkage studies suggested that a large share of familial early onset cases were instead associated with a locus on the long arm of chromosome 14 246 . ...
It is now well recognised that Alzheimer’s disease (AD) has a long presymptomatic period, with evidence of amyloid deposition being detected approximately 20 years before the onset of cognitive decline. This clinically silent period opens up a treatment window at a potentially more tractable stage of the disease. As we move towards AD prevention trials, there is a need for robust and sensitive methods that can detect and track disease progression, especially during this asymptomatic period. Familial Alzheimer’s disease (FAD) is an autosomal dominant condition that has many shared clinical, radiological, and neuropathological features with the more common sporadic form of disease. In addition, the age at onset in FAD is reasonably consistent between successive generations. Therefore, study of FAD mutation carriers provides an opportunity to prospectively characterise the sequence, and timings, of key pathological changes in AD. In addition, this reasonably pathologically pure form of AD provides a valuable opportunity to better understand the molecular drivers of disease onset. The studies presented in this thesis aim to facilitate the identification of biomarkers of FAD, with the overarching aim of identifying biomarkers of preclinical disease and/or key pathological processes. A multimodal approach is taken, with both presymptomatic and mildly symptomatic individuals being included. Chapter 1 introduces the background to the problems addressed in this thesis by providing an overview of early onset and familial Alzheimer’s disease and its preclinical period. Chapter 1 also outlines some of the key pathological drivers of AD onset, and specifically FAD. Chapter 2 then outlines the methodological approaches that are common to the different studies reported in this thesis. The first data chapter of this thesis (Chapter 3) performs data-driven modelling of cognitive performance in a clinically asymptomatic FAD cohort to demonstrate the sequence and timing of early cognitive change in FAD (Chapter 3). Following on from this the trajectory of a promising AD biomarker, plasma phospho-tau181 (p-tau181), 3 in FAD is explored (Chapter 4). This study, which was the first to examine the utility of plasma p-tau181 in FAD, showed that increased levels are detected in presymptomatic mutation carriers and that increases begin over a decade prior to estimated symptom onset. The ability of imaging measures of tau, specifically longitudinal tau positron emission topography (PET) scanning, to detect presymptomatic change is also examined (Chapter 5). In contrast to blood measures of p-tau181, this study did not find evidence of increased tau signal in presymptomatic mutation carriers, instead it showed tau burden in FAD increases in close temporal proximity to symptom onset. Finally, Chapter 6 investigates the influence of FAD genotype on plasma amyloid beta ratios. This study, the first to investigate inter-mutation differences in plasma amyloid beta profiles, showed evidence of marked inter-group differences as well as associations between the relative production of longer (≥ 42 amino acids length) amyloid beta peptides and the estimated timing of symptom onset in Presenilin1 mutation carriers. The thesis draws together these different approaches and discusses how they advance our understanding of the neurobiology of AD and their potential utility in both clinical assessment and presymptomatic therapeutic trials.
Over 55 million people globally are living with dementia and, by 2050, this number is projected to increase to 131 million. This poses immeasurable challenges for patients and their families and a significant threat to domestic and global economies. Given this public health crisis and disappointing results from disease-modifying trials, there has been a recent shift in focus toward primary and secondary prevention strategies. Approximately 40% of Alzheimer's disease (AD) cases, which is the most common form of dementia, may be prevented or at least delayed. Success of risk reduction studies through addressing modifiable risk factors, in addition to the failure of most drug trials, lends support for personalized multidomain interventions rather than a “one-size-fits-all” approach. Evolving evidence supports early intervention in at-risk patients using individualized interventions directed at modifiable risk factors. Comprehensive risk stratification can be informed by emerging principals of precision medicine, and include expanded clinical and family history, anthropometric measurements, blood biomarkers, neurocognitive evaluation, and genetic information. Risk stratification is key in differentiating subtypes of dementia and identifies targetable areas for intervention. This article reviews a clinical approach toward dementia risk stratification and evidence-based prevention strategies, with a primary focus on AD.
Background:
The 5XFAD model of Alzheimer's disease (AD) bearing five familial mutations of Alzheimer's disease on human APP and PSEN1 transgenes shows deposits of amyloid-β peptide (Aβ) as early as 2 months, while deficits in long-term memory can be detected at 4 months using the highly sensitive olfactory-dependent tests that we previously reported.
Objective:
Given that detecting early dysfunctions in AD prior to overt pathology is of major interest in the field, we sought to detect memory deficits at earlier stages of the disease in 3-month-old male 5XFAD mice.
Methods:
To this end, we used the Helico Maze, a behavioral task that was recently developed and patented. This device allows deeper analysis of learning and subcategories of hippocampal-dependent long-term memory using olfactory cues.
Results:
Eight male 5XFAD and 6 male wild-type (WT: C57Bl6 background) mice of 3 months of age were tested in the Helico Maze. The results demonstrated, for the first time, a starting deficit of pure reference long-term memory. Interestingly, memory impairment was clearly correlated with Aβ deposits in the hippocampus. While we also found significant differences in astrogliosis between 5XFAD and WT mice, this was not correlated with memory abilities.
Conclusion:
Our results underline the efficiency of this new olfactory-dependent behavioral task, which is easy to use, with a small cohort of mice. Using the Helico Maze may open new avenues to validate the efficacy of treatments that target early events related to the amyloid-dependent pathway of the disease and AD progression.
Both prevalence and incidence of AD increase steeply with advancing age in all populations investigated thus far. In general, women have higher prevalence and incidence; however, there are exceptions to this pattern. When considering only methodologically comparable surveys, and taking age into account, there are no major geographic differences in either prevalence or incidence. Data on incidence are unfortunately limited worldwide. There are no major time trends in either prevalence or incidence. The only definite risk factors for AD are age and familial aggregation. Putative risk factors are familial aggregation of Down’s syndrome, familial aggregation of Parkinson’s disease, late maternal age, head trauma, history of depression, and history of hypothyroidism. Cigarette smoking was found to be less common in the history of patients with AD than in that of controls. This association is probably spurious, however. Unfortunately, current knowledge about risk factors for AD does not justify the conduct of preventive trials or the introduction of large-scale interventions.
Background:
Polygenic hazard scores (PHS) estimate age-dependent genetic risk of late-onset Alzheimer's disease (AD), but there is limited information about the performance of PHS on real-world data where the population of interest differs from the model development population and part of the model genotypes are missing or need to be imputed.
Objective:
The aim of this study was to estimate age-dependent risk of late-onset AD using polygenic predictors in Nordic populations.
Methods:
We used Desikan PHS model, based on Cox proportional hazards assumption, to obtain age-dependent hazard scores for AD from individual genotypes in the Norwegian DemGene cohort (n = 2,772). We assessed the risk discrimination and calibration of Desikan model and extended it by adding new genotype markers (the Desikan Nordic model). Finally, we evaluated both Desikan and Desikan Nordic models in two independent Danish cohorts: The Copenhagen City Heart Study (CCHS) cohort (n = 7,643) and The Copenhagen General Population Study (CGPS) cohort (n = 10,886).
Results:
We showed a robust prediction efficiency of Desikan model in stratifying AD risk groups in Nordic populations, even when some of the model SNPs were missing or imputed. We attempted to improve Desikan PHS model by adding new SNPs to it, but we still achieved similar risk discrimination and calibration with the extended model.
Conclusion:
PHS modeling has the potential to guide the timing of treatment initiation based on individual risk profiles and can help enrich clinical trials with people at high risk to AD in Nordic populations.
The process of linkage analysis in humans requires a statistical framework in which various hypotheses about the linkage of a trait locus and marker locus can be considered in an observational study design. For the meiotic events of a parent to be scorable for linkage analysis, the parent must be heterozygous at both loci under consideration, i.e. a "double heterozygote". Computer simulation studies can be performed to assess the power of an available data set to detect linkage under an assumed genetic model prior to initiating a screen of the entire genome to detect linkage. Multipoint linkage analysis is a statistical technique using available genetic maps in which the order of markers and the distances between the markers are known. In contrast to the parametric methods discussed, nonparametric linkage analysis makes no assumptions about the underlying genetic model of the trait/disease being studied, only that the trait has some genetic component.
Early studies investigating the inheritance of schizophrenia estimated that between 66% to 93% of the variance in the aetiology is genetic. However, segregation analysis studying the patterns of affected individuals within families has not been able to establish a common mode of transmission. Nevertheless linkage analysis of marker data obtained from families is considered to be the most efficient and robust approach to studying the genetic aetiology of schizophrenia. Linkage studies in schizophrenia can be directed towards regions of the genome where there is a priori grounds for the presence of a susceptibility locus. This study investigated chromosome 5 after a report of an association (though weak) between a partial trisomy of chromosome 5q 11.2-q13.3 and schizophrenia.
Objectives: The genetic basis of schizophrenia has been clearly demonstrated through numerous well-conducted family, twin and adoption studies. However, the complex nature of the disorder has rendered it difficult to elucidate the mode of transmission and has made it difficult to identify the genes responsible for the disorder. The aim of this thesis is to employ recently developed statistical and molecular techniques to search for susceptibility genes. Hypothesis This thesis tests the hypothesis that linkage analysis can be successful in identifying susceptibility loci for schizophrenia in a sample of families specifically chosen for their large to medium size. From the outset, it was hypothesised that schizophrenia was a heterogenous disorder which was likely to have specific genetic factors operating within specific families. Methods: Given the complex nature of schizophrenia, genetic analysis of the disorder should employ samples and statistical methods capable of estimating several important variables such as lack of homogeneity (admixture), penetrance and the recombination fraction and with sufficient statistical power to test both single and polygenic hypotheses as well. The pedigree dataset for this study consisted of 23 Icelandic and English multiplex families which had been selected on the basis of stringent selection criteria, large to moderate family size, apparent dominant and unilineal transmission and the power to resolve heterogeneity of linkage. The hypothesis of linkage was tested using Restriction Fragment Length Polymorphisms (RFLPs), variable number tandem repeat markers (VNTRs) and highly polymorphic short tandem repeat (microsatellite) markers. The study can divided into three stages. For the first part of my study, I investigated the pericentric region on chromosome 5 implicated by Sherrington et al (1988) as the susceptibility locus by using a larger dataset of pedigrees and testing with more informative markers. Secondly, 1 expanded my study to other regions of the human genome. For this I used polymorphisms at or near the loci encoding candidate genes and favoured loci such as regions at which cytogenetic abnormalities had been reported as well as loci suggested by genome scans conducted by other groups. The third stage was a contribution to a very large collaborative effort initiated by the European Science Foundation. Linkage analyses were performed using parametric (lod score) and model-free linkage methods under the assumption of heterogeneity of linkage. In some instances the non-parametric sib-pair method was also used. The hypothesis of linkage was tested using two affection models assuming dominant transmission. Additionally, a range of chromosomal localisation techniques were employed to map glutamate receptor genes thought to be implicated in the aetiology of schizophrenia. Results: Lod and sib-pair linkage analyses showed that the dopamine receptor genes, D2, D3 and D5, and susceptibility loci on chromosomes 3p24-p26, 6p22-25, 11q, 15q, 22q22, Xp11 were not involved in the genetic susceptibility for schizophrenia in these particular families. An analysis of the pseudoautosomal region on the X chromosome was prompted by reports of sex concordance ratios in affected sib pairs and published positive linkage results. This region also proved negative for linkage. The analysis produced positive results on the pericentric region of chromosome 5 where a maximum lod of 1.19 was obtained and a maximised lod of 2.17 was obtained on chromosome 8p22-21. The glutamate receptor studies successfully localised the gene GRIN2D for the first time and confirmed localisations of the genes encoding the GRIN2A and GRIN2C receptor subtypes. Conclusion: The results are consistent with the hypothesis that linkage analysis may be able to detect susceptibility loci in schizophrenia but that such analysis in our sample failed to reach an acceptable level of statistical significance. When research on other samples is taken into account the hypotheses of linkage at these two loci gains support but are only fully confirmed in relation to the 8p22-21 locus. It seems likely that the failure to identify susceptibility loci unequivocally is due to locus heterogeneity. In order to overcome this problem, the chromosome 6, 8 and 22 data were combined into world wide collaborative samples. The results of this combined study helped to resolve the issue of replication to a certain extent but only in the case of the 8p22-21 locus because the traditional criterion for a confirmed linkage was reached in a replication sample. Efforts to detect linkage disequilibrium with schizophrenia on chromosome 8p should enable the hypothesis of linkage to be confirmed and for the aetiological mutations to be identified.
• In support of heterogeneity in Alzheimer's disease (AD), the existence of clinical and biologic subtypes has been claimed. We have investigated this claim by a statistical analysis of the relationships between the number of neurons in nucleus locus ceruleus (nLC), cortical levels of neurotransmitters, number of cortical plaques and tangles, and age. We separated AD patients into two groups: AD-1, with a less severe loss of nLC neurons; and AD-2, with a greater loss. The AD-2 cases were associated with less choline acetyltransferase activity, smaller concentrations of somatostatin and norepinephrine, and more plaques and tangles in the cerebral cortex. Although the mean age at death was less and the duration of dementia was greater in AD-2 patients than in AD-1 patients, the differences in these age-related variables were not significant. Further evidence of heterogeneity came from discriminant function analyses based on nLC neuronal counts and age at death. These findings, suggesting two subtypes of AD, suggest heterogeneity.
Early and accurate diagnosis of Alzheimer's disease (AD) has a major impact on the progress of research on dementia. To address the problems involved in diagnosing AD in its earliest stages, the National Institute on Aging, the American Association of Retired Persons, the National Institute of Neurological and Communicative Disorders and Stroke, and the National Institute of Mental Health jointly sponsored a workshop for planning research.
The purpose of the meeting was to identify the most important scientific research opportunities and the crucial clinical and technical issues that influence the progress of research on the diagnosis of AD. The 37 participants included some of the most knowledgeable and eminent scientists and physicians actively involved in the study of AD. The participants were divided among six panels representing the disciplines of neurochemistry, neuropathology, neuroradiology, neurology, neuropsychology, and psychiatry. Within each of the panels, participants discussed specific areas of research requiring further
Alzheimer's disease is a leading cause of morbidity and mortality among
the elderly. Several families have been described in which Alzheimer's
disease is caused by an autosomal dominant gene defect. The chromosomal
location of this defective gene has been discovered by using genetic
linkage to DNA markers on chromosome 21. The localization on chromosome
21 provides an explanation for the occurrence of Alzheimer's
disease--like pathology in Down syndrome. Isolation and characterization
of the gene at this locus may yield new insights into the nature of the
defect causing familial Alzheimer's disease and possibly, into the
etiology of all forms of Alzheimer's disease.
Alzheimer's disease is a devastating neurological disorder and the leading cause of dementia among the elderly. Recent studies have localized the gene for familial Alzheimer's disease to chromosome 21 in a series of early onset AD families (mean age of onset < 60). Familial late onset AD (mean age of onset > 60) is a more common clinical form of the disorder. Thirteen families with multiply affected Alzheimer's disease family members were identified and sampled. Ten of these families were of the late onset Alzheimer's disease type. Simulation studies were used to evaluate the usefulness of these pedigrees in linkage studies in familial Alzheimer's disease. Linkage studies undertaken to test the localization of both early onset and late onset Alzheimer's disease families to chromosome 21 failed to establish linkage and excluded linkage from a large portion of the region where the early onset Alzheimer's disease gene was localized. These findings suggest that more than one etiology may exist for familial Alzheimer's disease and indicate the need for continued screening of the genome in familial Alzheimer's disease families.
The DNA marker locus D21S13, localized in the 21q11.1-q21 region, has been closely linked to familial Alzheimer's disease. We constructed a physical map of 1.7 Mb around D21S13 using probes pGSM21 and pGSE9. The results indicated that pGSM21 contains recognition sites for at least three rare-cutting restriction enzymes. The clustering of rare-cutting restriction sites is indicative of the presence of an HTF (HpaII tiny fragment) island. Restriction site mapping and methylation analysis proved that pGSM21 contains a methylation-free HTF island. Furthermore, a cDNA correlate has been isolated confirming that pGSM21 is part of an expressed sequence. Today, the gene associated with pGSM21 is the gene closest to the centromere on the 21q arm.
Evidence from linkage studies suggests that familial Alzheimer disease (AD) can be caused by a defect in a gene on the proximal long arm of chromosome 21. We have constructed a physical map spanning 10 megabases of this region of the chromosome by means of pulsed-field gel electrophoresis and analysis of somatic cell hybrids. Our data have allowed us to establish the order of chromosome 21 loci--cen-(S16,S48)-S13-S46-S4-(S52,S110)-(S1,S1 1)--and are thus of immediate relevance both to multipoint linkage analysis in families affected by AD and for moving from this linkage to the isolation of the genetic defect. We have also been able to identify several CpG-rich sequences close to the four most centromeric loci, suggesting the location of genes in this region. These probes, which are all within 1.5 megabases of one another, are currently the markers most tightly linked to the AD locus. Genes identified in this region can therefore be considered as candidates for the disease locus.
We evaluated age at onset and lifetime risk for Alzheimer's disease (AD) in 70 kindreds with familial AD (designated FAD) composed of 541 affected and 1,066 unaffected offspring of demented parents who were identified retrospectively. Using a survival analysis method which takes into account affected persons with unknown onset ages and unaffected persons with unknown censoring ages, we found lifetime risk of AD among at-risk offspring by age 87 to be 64%. Analysis of age at onset among kindreds showed evidence for a bimodal distribution: in this sample, families with a mean onset age of less than 58 years were designated as having early-onset, while late-onset families had a mean onset age greater than 58 years. At-risk offspring in early-onset families had an estimated lifetime risk for dementia of 53%, which is significantly less than the risk of 86% that was estimated for offspring in late-onset families. Men and women in early-onset families had equivalent risk of dementia. In late-onset families, the risk to female offspring was somewhat higher than to male offspring but this difference was marginally significant. Lifetime risk of dementia in early-onset FAD kindreds is consistent with an autosomal dominant inheritance model. Our results may suggest that late-onset FAD has at least 2 etiologies; AD in some families may be transmitted as a dominant trait, whereas a proportion of cases in these and other late-onset families may be caused by other genetic or shared environmental factors.
Monoamine metabolites, biopterin, acetylcholinesterase (AChE) activity, and somatostatin-like immunoreactivity (SLI) were determined in the lumbar cerebrospinal fluid (CSF) of 24 patients with dementia of the Alzheimer type (DAT) without myoclonus or extrapyramidal signs, in 8 patients with DAT and myoclonus, and in 14 age-matched healthy control subjects. In patients with DAT with myoclonus as compared with both DAT patients without myoclonus and control subjects, the concentrations of homovanillic acid and biopterin were significantly decreased. 5-Hydroxyindoleacetic acid was significantly lower in patients with myoclonic DAT as compared to patients with nonmyoclonic DAT, but not significantly lower than in control subjects. CSF AChE and SLI were significantly reduced in patients with DAT with or without myoclonus, as compared with control subjects, but AChE and SLI were not significantly different between dementia groups. These results suggest that DAT patients with myoclonus represent a distinct clinical and neurochemical DAT subtype.