Diagnostic approaches to Alzheimer’s disease

Biochimica et Biophysica Acta 1502 (2000) 188^200 www.elsevier.com/locate/bba

Review

Diagnostic approaches to Alzheimer's disease Michael A. Boss * Athena Diagnostics, Inc, 377 Plantation Street, Worcester, MA 01605, USA Received 16 June 1999; received in revised form 1 December 1999; accepted 1 December 1999

Abstract The importance of obtaining an accurate and early diagnosis for Alzheimer's disease is now becoming recognized. Nonpharmacological as well as pharmacological therapies can be best initiated once a diagnosis is obtained. Biochemical markers to identify Alzheimer's disease have been sought for many years, with many candidates proposed. Recently criteria were established to evaluate putative diagnostic tests. Several biomarkers now show utility in identifying those with Alzheimer's disease. The ApoE e4 allele, while a risk factor rather than a deterministic gene, in the context of an individual with suspicion of AD has a positive predictive value of 94^98% and may come to have utility in predicting response to certain classes of pharmacological agents. Independent groups have shown that the markers in cerebrospinal fluid tau and Ab42 are, respectively, elevated and reduced in patients with AD versus other patient groups and that the lumbar puncture itself is usually well tolerated. For early-onset AD, sequencing presenilin 1 has come into use and the positive frequency is similar to that found in other genetic-based laboratory tests. ß 2000 Elsevier Science B.V. All rights reserved. Keywords: Diagnosis; Genetics ; Presenilin 1; Tau; Amyloid; ApoE; Biomarker ; Diagnostic marker; Alzheimer's disease

1. The importance of diagnosis While Alzheimer's disease (AD) is the most common cause of dementia, accounting for about 2/3 of all dementia cases, it is only one of about 70 di¡erent disorders that may cause dementia. Di¡erential diagnosis is di¤cult and conventional AD diagnoses have been reported as being wrong approximately 15% of the time from tertiary care centers. A number of these causes of dementia are currently reversible, including subdural hematomas, drug interaction, vitamin de¢ciencies, and low grade carbon monoxide poisoning. The more common non-AD causes of dementia are depression, parkinsonian, frontotemporal

* Fax: +1-508-752 7421; E-mail: [email protected]

and vascular dementias. Each of these conditions has unique treatment and management strategies, and therefore it is of key importance to distinguish between these di¡ering disorders as early and accurately as possible. Compared to other disease areas, the ¢eld of dementia oftentimes stimulates questions concerning the value of carrying out a diagnostic workup. It is sometimes said that since there is no cure or e¡ective therapy, that there is no point spending time and e¡ort in trying to identify the cause of the individual's disorder. A re¢nement of this argument is that, once potential treatable causes of dementia are ruledout, again there is no point in further classi¢cation. The argument is extended to comments that since an accurate diagnosis can only be made at autopsy, there is no reason to worry the person or the family by attaching a disease with a label that strikes fear

0925-4439 / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 5 - 4 4 3 9 ( 0 0 ) 0 0 0 4 4 - 2

BBADIS 61960 4-7-00

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

into many. Many of these types of comments are reminiscent of the approach taken until recently to patients with cancer. Indeed, until quite recently, it was possible for individuals in the UK to never know that they were dying of cancer. Clearly the way medicine is practised with respect to disclosure has changed. The last bastion of this approach is found in the dementia ¢eld. A study in a community-based retirement community showed that 80% of those who responded would prefer to know if they had AD [1]. The arguments concerning the pointlessness of carrying out a diagnostic workup are quite fallacious. In dementia, as in all other branches of medicine, the certainty of a diagnosis has an important impact on the management of the patient. While AD cannot be cured, there is symptomatic treatment available, and the ¢rst drugs for the temporary improvement of cognition and behavior are now licensed by the US Food and Drug Administration. Often overlooked, however, are the non-pharmacological interventions that can assist both the patient and the caregiver in tackling the day-to-day issues in living with this disorder. For example, quite simple techniques now in use can be used to help the patients improve their ability to dress themselves; and in some cases enable patients to dress themselves with only guidance. Behavioral therapies, while certainly not a cure, can have a quick practical impact as they have been shown to reduce many of the more disruptive behavior found in these patients, such as screaming, wandering or hitting [2]. Prior to the adoption of behavior management techniques, institutions would have typically attempted to control these behaviors by use of antipsychotic drugs or physical restraint, measures that increase the patient's agitation and may further worsen their condition. There are, therefore, signi¢cant non-pharmacological interventions possible that require as accurate a diagnosis as possible of the underlying disorder. Many times overlooked are the social issues concerned with the disease. Uniquely, dementia has a feature that is not replicated elsewhere. If a diagnosis is not provided until late in the disease process, the patient may be past the point at which they retain su¤cient cognitive ability to participate in their own medical treatment and planning. This can deprive the individual of including their desires and wishes in dealing with

189

the disease and place additional responsibilities on the caregiver. A diagnosis will enable the patients and their family to deal with legal and estate matters in a planned manner, to plan future care needs and address safety concerns, not the least of which in the US are the issues surrounding patients with AD continuing to drive. One of the key features of AD is the involvement of family members and typically a key individual as caregiver. Up to 3/4 of caregivers can develop signs of major depression, however, when involved in support services and training for managing behavioral problems, caregivers were found to respond favorably, and in one study delayed placing the patient in nursing homes by an average of one year. Some physicians indicate that unless a diagnosis is carried out and discussed with the caregiver, the caregiver may be reluctant to participate in support activities and they may not feel that it is appropriate to join in with such activities. This denies the caregiver the opportunity to come to terms with the disease, plan for the future, and learn best how to cope with the patient. One caregiver remarked that if she had known what was wrong with her father, she would not have gone ahead and bought a bigger house that was not suitable in which to look after him. This disease will have a major e¡ect on the lives of the caregivers, and they want to know what to expect as well as the patient. Intellectual decline is well recognized as a core feature of AD. What are often overlooked are the neuropsychiatric behavioral disturbances. It has been commented that AD is the most common psychiatric disorder after schizophrenia. These disturbances include personality disorders, delusions, hallucinations, mood abnormalities, disorders of sleep, change in appetite, altered sexual function, and disturbed psychomotor activity [3]. These neuropsychiatric disturbances of AD can be the most disabling consequences of the disease and are more troublesome to the family than the cognitive dysfunction. The neuropsychiatric disturbances are a typical feature of the disease, with 50% of AD patients have delusions. Delusional disturbances, for example, are often treatable with conventional neuroleptic agents and the therapy may well decrease the associated aggression and agitation, improving the quality of life for the patient and their caregiver. The di¡ering causes of

BBADIS 61960 4-7-00

190

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

dementia have a di¡erent frequency distribution of neuropsychiatric disturbances. It is crucial, not only to treat these disturbances, but also to know based on the speci¢c disease, the most likely panoply of disturbances that will be encountered and to provide training for the caregiver in management of the patient. As with any disorder, prior to the use of complex pharmacological interventions, it is key to establish the most accurate diagnosis possible as well as identify any concurrent illnesses. While it is conventionally considered that the `gold standard' for a diagnosis of AD are the neuropathology ¢ndings at autopsy, there remain di¡erences of opinion as to how to interpret the neuropathological data in the event of concomitant diseases, a not uncommon feature of AD. Two drugs, both acetylcholinesterase inhibitors, are currently approved by FDA for the direct treatment of AD. These agents, however have side e¡ects and the indiscriminate use of these drugs in all dementia patients is not recommended [4]. Use in patients who do not have AD results in patients being treated for a disease that they do not have, and could mean that they do not receive treatment for the disease that they do have. 2. Conventional diagnostic approaches While the diagnostic workup of a patient according to NIH recommended guidelines may cost approximately $2000, this is dwarfed by the other costs of the disease. It is estimated that in the US the cost of caring for a patient by their family is $12 000 a year and in a nursing facility in California the average cost is reported at $47 000 a year, of which the family typically pays about half. It is estimated that within 50 years, if unchecked, there may be 14 million individuals in the US who may be a¡ected with AD and will require medical care and possibly institutionalization [5]. Indeed the frequency of unrecognized dementia could well be of staggering proportions. The Honolulu-Asia aging study reported that as many as 21% of families of persons considered to have dementia did not recognize the presence of disease [6] and a staggering 52% of family members did not recognize the presence of very mild dementia in the patient. In total some 60% of the subjects identi¢ed as having dementia were either not recognized

by their family as such or had not been medically evaluated for dementia. Studies on the prevalence of dementia in East Boston [7] also support the concept that dementia is even more prevalent than typically considered. Were the Honolulu-Asia aging study frequency of unrecognized dementia broadly true, this would imply that there are about 1.8 million people currently in the US with unrecognized dementia in addition to the 2 to 4 million cases of dementia currently acknowledged. There is a great need, therefore, for ways to enhance our abilities to identify AD as early and as accurately as possible. Practitioners are now recognizing the bene¢ts from an early and accurate diagnosis. However, diagnosis at the earlier stages of the disease is likely to be less accurate than after prolonged follow-up. Indeed, one of the tenets of the criteria used in diagnosis is the measurement of a gradual cognitive decline over an extended period. There are two current roles for differential diagnosis. The ¢rst is to identify the cause of the dementing disorder, and to identify whether the individual is su¡ering from AD or one of the other 70 causes of dementia or a combination of dementing disorders. Equally problematical is the second di¡erential, which is to identify whether the individual actually has dementia versus the currently considered normal e¡ects of aging. The distinction between normal aging, so-called age-related memory impairment, and dementia is a topic of active current research. The advent of both biological based diagnostic tools, and the forecast of more and better therapeutic interventions will drive the ¢eld to establish clearer distinctions. Indeed, the entire concept of age-related memory impairment may come to be seen as analogous to the e¡ect of a lack of estrogen in post-menopausal women. The development of osteoporosis, another example of a late-onset disease, may be normal but it is neither desirable nor inevitable. There are both pharmacological and behavioral interventions possible to reduce the risk of osteoporosis. In a similar way, the ¢eld of dementia may develop approaches that are used from an early age to reduce the probability of developing Alzheimer's disease, coupled with identifying high-risk individuals and initiating possible pharmacological interventions prior to the development of any symptoms. The management of the populations at various age ranges and risk categories would involve widespread screens

BBADIS 61960 4-7-00

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

and interventions based on the individual's risk pro¢le and age. With the onset of AD symptoms, a di¡erent set of management tools and pharmacological interventions as well as monitoring studies would be initiated. In this way, the vision of AD over the next 5^10 years might come to resemble those re¢ned for the management of disorders such as heart disease. The clinical diagnosis of AD requires a history from the patient and ideally a relative or other caregiver and results of mental status testing, physical and neurological examinations, a panel of conventional laboratory tests, and a neuroimaging study to rule out other causes of dementia. The workup may include additional laboratory tests in atypical cases and more detailed psychometric testing and follow-up to increase the con¢dence of the diagnosis, especially when a patient presents with mild dementia [8]. The current guidelines recognize an element of uncertainty in the clinical diagnosis with the use of the terms `possible' or `probable', with de¢nite AD reserved for autopsy proven cases. Studies of the accuracy of diagnosis of patients who were evaluated at dementia specialty research centers and followed through to autopsy, report rates of agreement of 85^90%. These data should, however, be considered in context. The patients received extensive workups and longitudinal monitoring of the cognitive decline. However, the resources applied to carry out these evaluations exceed those available to the typical clinical practice. With approximately 500 000 new cases of dementia a year in the US, the forecast of signi¢cant growth in these numbers with the aging population, and the decrease in many other diseases of aging, only very few of these individuals will be fortunate enough to be evaluated in a specialized referral center. Currently approximately 200 new patients a year are evaluated at each Alzheimer Disease Research Center. Thus with 30 or so centers in the US, some 6000 patients a year bene¢t from these referral centers, approximately 1% of the annual caseload. Diagnosis at early stages of disease is likely to be less accurate than after prolonged follow-up. With the advent of symptomatic treatment for cognitive symptoms of AD and as drugs are developed to slow the progression of AD, very early diagnosis is important to de¢ne the earliest window of opportunity to apply such treatments. At the earliest

191

stages of the disease, it is particularly di¤cult to document a history of progressive cognitive and functional decline or to demonstrate de¢cits on cognitive testing. This can raise questions of whether the symptoms represent age-associated changes or AD. The increasing prevalence of AD, the limited number of specialty dementia clinics, the increased necessity to make an accurate diagnosis to dictate therapy all further pressure the traditional diagnostic process. The advent of biomarkers may help to resolve these dilemmas. 3. Criteria for assessing new diagnostic markers The search for biological markers of dementia and AD has been fraught with misunderstandings and unrealistic expectations against which any putative marker will fail [9]. There are a number of di¡erent potential uses for biomarkers in AD evaluation, and each use could involve a di¡erent marker or set of markers. First is the use of a marker to distinguish AD from other causes of dementia. Second is to distinguish dementia from the non-pathological effects of aging. Third is to monitor the progress of the disease after clinical symptoms become apparent. Fourth will be to have a surrogate to monitor the e¤cacy of the forthcoming therapies for AD. Fifth there is need for markers which are risk factors for AD and ¢nally identify both the earliest biological changes occurring in the brain and other changes that occur as the disease progresses. The concept that any one marker will ful¢ll all these very di¡ering requirements to a high degree of sensitivity and speci¢city is na|«ve. Any individual marker needs to be assessed by sensitivity, speci¢city, reliability and validity for the type of clinical situation to which it is meant to appertain. A marker that is poor at distinguishing AD from other causes of dementia, could nevertheless be an excellent marker to monitoring the progression of the disease process or the response to therapy. These potential di¡ering uses of biomarkers need to be kept in mind as information emerges. Curiously, there appears to be an arti¢cial issue a£oat, which is the idea that a biomarker or series of biomarkers will be competitive with the insights of a trained physician. The more likely scenario is that the diagnosis and therapy of dementia will come to

BBADIS 61960 4-7-00

192

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

resemble other late-onset disorders such as the prevention and control of coronary artery disease. The clinical evaluation and biological markers will go hand in hand to evaluate risk, monitor developments, plan for and monitor interventions. There are a number of reasons to consider biological markers for AD, which include strengthening the certainty of the clinical diagnosis, distinguishing AD from other causes of dementia, and providing a way to stage the disease and the level of disease activity and rate of progression. In the future, therapies may be best monitored by use of surrogate biological markers. Such an approach may complement the evaluation of the clinical assessment and could help streamline the evaluation of future drugs [8]. Biochemical markers for AD have been sought for many years. The ¢eld is replete with publications of single studies that could not later be con¢rmed. The report in 1994 by Scinto et al. on the development of a pupil dilation test for AD, is one such ¢nding that at the time stimulated great excitement [10]. A subsequent series of 18 di¡erent studies failed to identify a useful correlation between pupil dilation with tropicamide and AD [11]. As will be discussed below other approaches have proven robust and reproducible in many laboratories. With the emergence of the ¢rst biological markers that may stand up to scrutiny, a working group from the National Institute of Aging under the auspices of the Alzheimer's Association was set up to identify the criteria by which a potential biological marker should be evaluated and whether any currently identi¢ed markers met the de¢ned criteria. The NIA/AA group identi¢ed a series of ideal criteria for a putative diagnostic test and that any proposed marker should include as many of these ideal features as possible. In this way, the development of a standardized approach will aid the development and evaluation of putative biological markers. Such an approach will provide a framework by which new discoveries will be measured to guide further work and discourage premature claims of diagnostic utility based on single studies or those carried out by only one group. The criteria identi¢ed were: ability to detect a fundamental feature of AD's neuropathology; validated in neuropathologically con¢rmed AD cases; able to identify AD early in the course of disease and distinguish AD from other dementias;

reliable; non-invasive; simple to perform; and inexpensive. In addition, the group de¢ned a recommended ¢ve steps for establishing a biomarker (see Table 1). Using these criteria the Working Group concluded that in early-onset familial AD that it was appropriate to search for mutations in the presenilin 1, presenilin 2 and amyloid precursor protein genes. In late-onset and sporadic AD it was concluded that detecting an apolipoprotein E O4 allele can add con¢dence to the clinical diagnosis. Among a number of markers evaluated, cerebrospinal £uid assays measuring levels of beta amyloid 1^42 (AL42 ) and tau protein came closest to ful¢lling the criteria for a useful biomarker. 4. ApoE in Alzheimer's disease The ApoE O4 allele is a risk factor for AD rather than a deterministic gene. Nevertheless, most cases of AD with onset between the ages of 50^60 have been attributed to the inheritance of an O4 allele. The ¢rst link between ApoE and AD was reported in 1993 by Corder et al. [13]. This study identi¢ed that the risk for AD increased from 20 to 90% and that the average age of disease onset decreased with an increasing number of ApoE O4 alleles. The mean age of onset of AD was 68 years in those with two O4 alleles, 77 years with one O4 allele in contrast to 85 years in those with no O4 alleles. The presence of two O4 alleles was virtually certain to lead to AD by age 80. Table 1 Recommended steps in the process of establishing a biomarker [12] 1.

2. 3.

4.

5.

There should be at least two independent studies that specify the biomarker's sensitivity, speci¢city, and positive and negative predictive values. Sensitivity and speci¢city should be no less than 80%; positive predictive value should approach 90%. The studies should be well powered, conducted by investigators with expertise to conduct such studies, and the results published in peer-reviewed journals. The studies should specify type of control subjects, including normal subjects and those with a dementing illness but not AD. Once a marker is accepted, follow-up data should be collected and disseminated to monitor its accuracy and diagnostic value.

BBADIS 61960 4-7-00

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

These original observations have been con¢rmed by many studies worldwide, and the in£uence of ApoE in the disease process is now universally accepted. The O4 allele is considered the strongest risk factor of late-onset AD next to age, whereas the O2 allele may be protective or to delay disease onset, which may amount to the same thing. The ApoE genotype is perhaps the most important biological marker identi¢ed for susceptibility to AD to date and accounts for some 40^60% of its genetic component [14]. Unlike the deterministic mutations in genes causative of disease, the ApoE alleles are distributed in all ethnic groups with variations in relative allele frequencies. While no prediction of whether AD will occur in an individual based on their allelic makeup is possible, similarly no ApoE genotype provides an absence of risk [15]. It is important to note that not everyone with an O4 allele will get AD and not everyone who gets AD will have an ApoE O4 allele. In autopsy series over half of the patients with a con¢rmed diagnosis of AD had an O4 allele. Testing for ApoE has come to be accepted as a valuable adjunct in the evaluation of a patient with dementia. The ¢nding of an ApoE O4 allele in context with a clinical workup and suspicion of AD has a demonstrated positive predictive value of 94^98%. The greatest utility of ApoE analysis is early in the course of the disease when an accurate diagnosis can be particularly challenging. Two large studies in particular have helped to clarify the initial observations with ApoE. The ¢rst was a meta-analysis carried out by Farrer et al. [16] which considered 40 di¡erent studies involving 5930 patients with probable or de¢nite AD, and 8607 controls without AD from a database of more than 15 000 patients and control subjects. The results showed that the ApoE O4 allele represents a major risk factor for AD in all the ethnic groups studied, across all ages from 40 to 90 years and in both men and women. The second major study that con¢rmed the value of ApoE analysis in the evaluation of a person with dementia utilized autopsy proven diagnoses and was reported by Mayeux and colleagues [17]. This very large study followed the earlier publication of a number of smaller autopsy-based studies where the speci¢city for the presence of an O4 allele and autopsy con¢rmation of AD was found to be 100% [18^20].

193

Given that the O4 allele is a risk factor and not a causative genetic alteration, these ¢ndings were particularly surprising. Mayeux and colleagues, in a study designed to lay to rest the issue of whether ApoE O4 analysis was useful or not in the diagnostic workup of a patient with dementia, reported on an autopsy-con¢rmed study of 2188 patients from 26 of the AD research centers of the NIH. They found that the sensitivity of clinical diagnosis alone was 93% at these specialist centers. However, the speci¢city was found to be only 55%. The sequential use of ApoE analysis enhanced the speci¢city of diagnosis from 55 to 84%, an increase in speci¢city of some 53% from clinical evaluation alone. Adding ApoE analysis increased the positive predictive value of the workup from 89.6 to 94.2%. Used alone, however, as might be expected of a marker of trait that is a risk factor, the sensitivity and speci¢city of ApoE was low. Thus the use of ApoE testing was unambiguously con¢rmed as clinically relevant in patients with dementia and suspected AD. The ¢nding of an O4 allele in the context of a dementing disorder has now become an accepted part of the standard workup of a patient with dementia. The presence of an ApoE O4 allele in a person with dementia strongly indicates the presence of AD as causative of or contributory to the dementing disorder. The absence of an O4 allele, however, does not rule out the possibility of AD. ApoE genotyping is an analysis with the unusual characteristic of having a strong positive predictive value but a weak negative predictive value. ApoE allelic analysis has also been investigated in the novel area of pharmacogenomics. For probably the ¢rst time, a genetic marker was evaluated to identify whether a particular allele caused a heightened or diminished response to a pharmacological agent. Farlow et al. [21] reported on ApoE genotype in relationship to therapy with tacrine. This is an agent with a signi¢cant side e¡ect pro¢le, and only 25^50% of treated patients show a bene¢cial clinical response. This initial study gave uncertain results. Patients on tacrine and with an O4 allele were found to be less responsive to therapy than those with other allele types as a group. However, some individual patients with an O4 allele did in fact respond very well to the therapy. Curiously, the patients in the untreated group who had an O4 allele declined more slowly than those patients in the untreated

BBADIS 61960 4-7-00

194

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

group who had an O2 or O3 allele. This study suggested a relationship between ApoE genotype and therapeutic response to cholinesterase therapy, but further studies are required to clarify some of the confusing results from this study. In particular, it would be interesting to determine the relationship between ApoE genotype and response to donepezil. Nevertheless, the Farlow study lay the way for stratifying patients in drug trials by ApoE genotype and potentially other genetic markers, and seeking way to use genetic analysis to predict response or non-response to a particular pharmacologic agent. 5. Tau Progress in identifying biomarkers of AD recently has focused on analytes in cerebrospinal £uid (CSF) that are based on the pathologically altered proteins found in the brains of patients with AD. Outside of

the neurological community there remains a concern surrounding the drawing of CSF despite the proven utility and safety of the procedure. Lumbar puncture is a minimally invasive outpatient procedure that is in routine use. Serious complications of lumbar puncture are extremely rare and the procedure is usually well tolerated. The most signi¢cant adverse e¡ect that is normally associated with lumbar puncture is postlumbar puncture headache, which is estimated to occur in 10% of cases in the general population [22]. This is thought to be due to a lowering of CSF pressure after the lumbar puncture because of persistent leakage at the puncture site. More recently the use of small-caliber needles has reduced this risk. Recent studies on postlumbar puncture headache show a low incidence of headache and that the procedure is well tolerated. A large study of 395 elderly patients found that only 2% developed a headache [23] and a second study found that 92% of elderly patients agreed to a second lumbar punc-

Table 2 Tau levels in AD versus control groups from 2339 patients Study

AD

Seubert [22] Galasko [25] Kanai [26] Motter [27] Vigo-Pilfrey [28] Munroe [29] Arai [30] Galasko [31] Riemenschneider [32] Tato [33] Jensen [34] Blomberg [35] Rosler [36] Hock [37] Mori [38] Riemenschneider [39] Rosler [40] Blennow [41] Skoog [42] Isoe [43] Vandermeeren [24] Arai [44] Vanderstichele [45] Total

76 u 82 u 45 u 37 u 71 u 24 u 70 u 36 u 22 u 23 u 21 u 18 u 20 u 19 u 14 u 39 u 16 u 44 u 11 u 9u 27 u 87 u 81 u 892

VAD

Depress Frontal

PD

32 H 25 H 26 H 96 Hu 9H 3 Hu

36 H

12 H

17 u 13 u

9H

11 H 18 H

11 H 10 H

19 Hu 21 Hu 118

Non-AD

50

2H 11 H

4H 15

12 Hu

15 H

229

AMCON

53 76 Hu 60 24 20 26 13 H 14 19 14 19 23 22 9H 12 1u

2u

13 H 17 Hu 30

Other

36 30 10 31 35

YCON

H H H H H H H H H H H

8H

H

18 H

H H H H H

16 H

110 Hu 39 u 22 H 15 H 250 485

51 H

93

NCON

Total

129 218 8H 77 89 59 H 181 77 185 10 H 69 44 82 31 Hu 82 36 24 Hu 85 37 63 83 20 Hu 73 128 61 9 207 25 H 224 100 177 2339

AD, Alzheimer's disease; VAD, vascular dementia; Depress, depression; Frontal, frontal lobe dementia; PD, Parkinson's disease; Non-AD, other dementing illness (not AD); Other, varied disease states; AMCON, age-matched controls; YCON, younger controls; NCON, neurological controls (not demented); u, tau elevated; H, baseline; Hu, slight tau elevation.

BBADIS 61960 4-7-00

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

195

Table 3 Correlation of tau and demographics Study

Cog. score

Age

Seubert [22] Galasko [31] Kanai [26] Motter [27] Vigo-Pilfrey [28] Munroe [29] Arai [30] Galasko [25] Riemenschneider [32] Tato [33] Jensen [34] Blomberg [35] Rosler [36] Hock [37] Mori [38] Riemenschneider [39] Rosler [40] Blennow [41] Skoog [42] Isoe [43]

none

none

none insigni¢cant none insigni¢cant none none signi¢cant

none insigni¢cant insigni¢cant

Arai [44] Vandermeeren [24] Vanderstichele [45]

slight increase increase

Duration of AD

none none

Age of onset

increased

insigni¢cant none

insigni¢cant

Gender

AD genotype

none

(ApoE) none

none

(ApoE) insig.

none

slight increase

none

(APP) increase (ApoE) increase

insigni¢cant

signi¢cant

insigni¢cant

increased, then decreased

increased increased, then decreased none increase

(ApoE) none

(ApoE) none

none

none

(ApoE, PS-1, ACT) none none

Information re£ects terminology used by the authors. Please see individual studies for further information.

ture, demonstrating the general acceptability of the procedure. While not trivial, lumbar punctures in elderly demented patients are a relatively straightforward technique, and the potential information gathered from CSF analysis outweighs the minimal discomfort involved. The correlation between the abundance of neuro¢brillary lesions and the clinical measurements of dementia stimulated the search to determine whether this correlation could be measured in CSF. The initial breakthrough came in 1993 when Vandermeeren and colleagues [24], using a very sensitive assay that could measure tau down to 5 pg/ml, found that CSF tau levels were signi¢cantly elevated in patients with AD compared to healthy or other disease controls. Tau is a microtubule-associated protein and the primary constituent of tangles. Subsequent to this initial study a surge of recent studies by independent groups con¢rms that tau levels are increased in the CSF of patients with AD, compared to age-matched control subjects.

Over 23 studies have been published in peer-reviewed journals from groups around the world showing that tau is elevated in the CSF of AD patients when compared to either age-related controls or patients with other neurological diseases. This work represents testing tau on 2339 subjects involving 892 patients with AD (Table 2). Because of the differences in assays and clinical criteria used, it is particularly surprising that the results are similar and that every group found an increase in tau levels in AD patients which is both speci¢c and reasonably sensitive, with typical measures of 70 and 82% respectively when compared to non-demented control subjects. The AD patients have also been compared to patients with a series of other disorders and other dementias including: vascular dementia, depression, di¡use Lewy body disease, Pick's disease, Parkinson's disease, and Creutzfeld^Jakob disease. These represent a total of some 213 subjects, and no disease showed a consistent elevation of tau. In general the studies show that there is no correlation between

BBADIS 61960 4-7-00

196

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

levels of tau and age or gender (Table 3). Further, there was in general no correlation between levels of tau and gender, cognitive scores, or duration of disease. Thus elevation of tau is both a sensitive and speci¢c marker for AD. One concern with the measurement of tau levels was that the measurement may merely represent a proxy for the detection of neuronal cell loss. However, absence of elevated tau in a variety of other degenerative disorders including HIV-associated neurocognitive disorders [46] indicates that the elevation of tau is not simply correlated with loss of cells. An exception to this ¢nding is the enormous spike of tau found in the few months following a stroke, which are typically far higher than those seen in AD. The absence of elevated levels of tau in 485 agematched controls is of particular interest. The absence of elevated tau from so many individuals, some of whom, given the age pro¢le, will soon become demented is in contrast to the ¢nding that tau is elevated in many of the most mildly demented patients that can be reliably identi¢ed clinically [31]. These observations imply that the gross observation of elevated tau in CSF is a relatively late occurrence in the disease process and coincident with the stage of disease at which clinical onset of symptoms becomes apparent. 6. AL L42 The confusion produced by the initial observations of the absence of a correlation between amyloid levels in CSF and AD were resolved by Motter et al.

[27] who measured both total AL levels, which include the predominantly present species AL40 , and the amyloidogenic fragment AL42 which comprises about 5% of total soluble AL. Total AL levels did not di¡er from AD patients and age-related healthy controls or disease controls. However, AL42 levels were signi¢cantly lower in AD patients versus either control group. These initial ¢ndings were later extended as well as con¢rmed by independent groups. Seven studies on AL42 in CSF show a consistent ¢nding that AL42 is reduced in the CSF in the 375 patients studied in comparison to 206 age-matched controls (Table 4). These studies compared patients with AD to those with other disorders including vascular, frontotemporal and Parkinson's dementia. As with tau, levels of AL42 are not correlated with age or age of onset of disease. Thus lower levels of AL42 are not part of the normal aging process but a direct measurement of a pathological alteration. Similarly there is no correlation of AL42 with gender, or cognitive test scores. One of the criteria for a potentially useful diagnostic test is a biologically plausible relationship of the marker to the pathogenesis of the disease. Increased production and or deposition of AL42 is stimulated by all of the known genetic causes of early-onset AD as well as the presence of an ApoE O4 allele, and deposits of AL42 as di¡use plaques are among the earliest detectable neuropathological abnormalities of AD. Therefore the AL42 peptide is of core importance to the disease pathology and the measurement of CSF AL42 is causally related to the disease process and not merely a sequelae of the disease. The decrease of AL42 in CSF is possibly due to a decrease in the levels of the peptide in the brain

Table 4 AL42 levels in AD versus control groups from 726 patients Study

AD

Seubert [22] Galasko [25] Kanai [26] Motter [27] Tamoka [47] Vanderstichele [45] Ida [48] Total

76 s 82 s 45 s 37 s 20 s 81 s 34 s 375

VAD

Depress

Frontal

PD

Non-AD Other 76 Hs

32 H 4s 4

14 s 46

76

53 60 24 20 34 15

H H H H H H

206

AMCON NCON

8H

11 H 19

Total 129 218 77 89 54 100 59 726

AD, Alzheimer's disease; VAD, vascular dementia; Depress, depression; Frontal, frontal lobe dementia; PD, Parkinson's disease; Non-AD, other dementing illness (not AD); Other, varied disease states; AMCON, age-matched controls; YCON, younger controls; NCON, neurological controls (not demented); s, AL42 decreased; H, baseline; Hs, slight AL42 decreased.

BBADIS 61960 4-7-00

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

interstitial £uid as the AL42 becomes increasingly insoluble and forms the plaque deposits. In collaboration with six AD research centers, our group extended its initial observations in a clinical study involving 82 patients with probable AD, including 24 with very mild dementia, 60 cognitively normal elderly controls and 74 patients with a range of neurological disorders including dementia [25]. Levels of AL42 and tau were analyzed and compared across the groups and compared to ApoE genotype using multiple linear regression. Across the age range studied, there was no correlation between age and CSF levels of these analytes. Consistent with the other studies, levels of AL42 were signi¢cantly lower and levels of tau were signi¢cantly higher in the AD group than in either control group. In the AD group, the level of AL42 was correlated with the presence of an O4 allele, with the levels being lowest in those patients homozygous for the O4 allele. Regardless of whether subjects had none, one or two O4 alleles, the AD group showed lower AL42 mean levels than the normal control group. Using both tau and AL42 levels, at a cuto¡ for speci¢city de¢ned as at least 90%, the combined marker analysis provided a sensitivity of 77% for AD. Kanai and colleagues from three institutes in Japan performed a large study of 236 patients [26] and con¢rmed the combined utility of tau and AL42 . This study combined use of tau and AL42 provided a diagnostic sensitivity of 71% and a speci¢city of 83%. Intriguingly, given that patients with AD have low levels but not high levels of AL42 , while age-related controls have relatively higher levels of AL42 but may have low levels of AL42 , this could imply that the lowering of AL42 could signify an early stage of the disease process. Clearly speculative, however, is that the measurement of a reduction of AL42 in CSF could be an early indicator of disease. Once e¡ective AD therapies are developed, if prospective studies show that reduction in CSF AL42 is an early indicator of the buildup of pathology, the monitoring of AL42 CSF levels might be a key marker in the monitoring of the population at risk. CSF levels of tau and AL42 have each shown good correlation with AD independently. The combination together of these markers enhances the value of each marker independently. In patients suspected of having AD, but in whom the clinical diagnosis is uncer-

197

tain, CSF tau and AL42 can be used to strengthen the diagnosis. When more than one potential cause of dementia is suspected, these CSF markers can be used to help establish that AD is a contributory factor. 7. Disease causative genes The major genetic causes of early-onset AD are now clearly identi¢ed. These are classical autosomal dominant traits with virtually complete penetrance and resemble in their e¡ect other conventional adult-onset disease causative genetic markers such as Huntington's disease. The characteristics of the disease are normally the same in those with earlyonset AD compared to the symptoms and progression found in late-onset AD. However, a number of cases have been described indicating that the disease phenotype of early-onset AD can be broader than conventional symptoms for AD. Genetic screening is indicated for individuals showing a family history combined with dementia combined with atypical neurodegenerative phenotypes more typical of other disorders including diseases such as Pick's disease, cerebral hemorrhage, or spastic paraparesis [49]. Mutations in the presenilin 1 gene are the most common de¢ned cause of early-onset familial AD, and are estimated to account for 30^50% of all cases of early-onset autosomal dominant familial disease [50]. The other genetic causes of early-onset familial AD, mutations in the L amyloid precursor protein and presenilin 2 (PS2) are considered restricted to a few families and are not considered at this time of general utility outside of a research setting. Our laboratory developed and set up a diagnostic test for early-onset AD based on sequencing the entire 1401 base coding region of the presenilin 1 gene. Some authors have indicated that the frequency of early-onset familial AD may be much less frequent than is generally acknowledged. Our data show that out of the ¢rst 195 patients sequenced, 32 (16%) had a mutation. Given that some of these patients may be related, by counting only separate mutations, we found that 18 (9%) di¡erent mutations were identi¢ed, each representing a di¡erent lineage. This frequency of mutation identi¢cation (9^16%) in a relatively unselected group is similar to the positive

BBADIS 61960 4-7-00

198

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

frequency found in other genetic-based clinical laboratory tests, and gives reason to believe that mutations in PS1 are responsible for a signi¢cant number of cases of dementia in the population. Analysis of PS1 for mutations can be used for either diagnostic or prognostic information. Once an a¡ected individual in a family is identi¢ed, the other members of the family have the opportunity to be tested prognostically should that be of interest to them. As a deterministic genetic marker for an adult-onset disorder, the issues surrounding testing for mutations in PS1 are similar to those for other such genetically causative disorders and have been well de¢ned. These relate primarily to counseling for those seeking presymptomatic testing, and the moral and ethical issues surrounding testing of children and prenatal testing. More recently mutations in tau have been shown to cause a form of frontotemporal dementia, FTDP17. The phenotypes caused by these mutations are still being de¢ned, as is the proportion of frontotemporal dementia that is caused by mutations in the tau gene. Clearly the onward march of genetics is beginning to make inroads into the tortuous ¢eld of dementia. As in other disease areas, it is likely that the very de¢nitions of the diseases themselves will come to be based on the speci¢c underlying genetic alterations.

[8] [9]

[10]

[11] [12]

[13]

[14]

[15]

[16]

References [1] S. Holyroyd, D.G. Snustad, Z.L. Chalifoux, Attitudes of older adults' on being told the diagnosis of Alzheimer's disease, J. Am. Geriatr. Soc. 44 (1996) 400^403. [2] M. Barinaga, Alzheimer's treatments that work now, Science 282 (1998) 1030^1032. [3] J. Cummings, Recognition and management of behavioral disorders in Alzheimer's disease and other dementias, Am. Acad. Neurol. 143 (1995) 143^163. [4] S. Lovestone, N. Graham, R. Howard, Guidelines on drug treatments for Alzheimer's disease, Lancet 350 (1997) 232. [5] Z. Khachaturian, Editorial, Neurobiol. Aging 19 (1998) 107. [6] G.W. Ross, R.D. Abbott, H. Petrovitch, K.H. Masaki, C. Murdaugh, C. Trockman, D. Curb, L.R. White, Frequency and Characteristics of silent dementia among elderly Japanese-American men, J. Am. Med. Assoc. 277 (1997) 800^ 805. [7] D.A. Evans, H.H. Funkenstein, M.S. Albert, P.A. Scherr, N.R. Cook, M.J. Chown, L.E. Hevert, C.H. Hennekens,

[17]

[18]

[19]

[20]

[21]

J.O. Taylor, Prevalence of Alzheimer's disease in a community population of older persons, J. Am. Med. Assoc. 262 (1989) 2551^2556. D. Galasko, CSF Tau and AL42: Logical Biomarkers for Alzheimer's disease?, Neurobiol. Aging 19 (1998) 117^119. N.L. Foster, The development of biological markers for the diagnosis of Alzheimer's disease, Neurobiol. Aging 19 (1998) 127^129. L.F. Scinto, K.R. Da¡ner, D. Dressler, B.I. Ransil, D. Rentz, S. Weintraub, M. Mesulam, H. Potter, A potential noninvasive neurobiological test for Alzheimer's disease, Science 266 (1994) 1051^1054. R.H. Kardon, Drop the Alzheimer's drop test, Neurology 50 (1998) 588^591. The Ronald and Nancy Reagan Research Institute of the Alzheimer's Association and the National Institute on aging Working Group, Consensus Report of the Working Group on `Molecular and biochemical markers of Alzheimer's disease', Neurobiology 19 (1998) 109^116. E.H. Corder, A.M. Saunders, W.J. Strittmatter, D.E. Schmechel, P.C. Gaskell, G.W. Small, A.D. Roses, J.L. Haines, M.A. Pericak-Vance, Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer's disease in late onset families, Science 261 (1993) 921^923. J. Nalbantoglu, B.M. Gil¢x, P. Bertrand, Y. Robitaille, S. Gauthier, D.S. Rosenblatt, J. Poirier, Predictive value of apolipoprotein E genotyping in Alzheimer's disease: Results of an autopsy series and an analysis of several combined studies, Ann. Neurol. 36 (1994) 889^895. A.D. Roses, A model for susceptibility polymorphisms for complex diseases: apolipoprotein E and Alzheimer disease, Neurogenetics 1 (1997) 3^11. L.A. Farrar, L.A. Cupples, J.L. Haines, B. Hyman, W.A. Kukull, R. Mayeux, R.H. Myers, M.A. Pericak-Vance, N. Risch, C.M. van Duijn, E¡ects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer's disease, J. Am. Med. Assoc. 278 (1997) 1349^1356. R. Mayeux, A.M. Saunders, S. Shea, S. Mirra, D. Evans, A.D. Roses, B.T. Hyman, B. Crain, M. Tang, C.H. Phelps, Utility of the apolipoprotein E genotype in the diagnosis of Alzheimer's disease, N. Engl. J. Med. 338 (1998) 506^511. B.A. Kakulas, S.D. Wilton, V.A. Fabian, T.M. Jones, Apolipoprotein-E genotyping in diagnosis of Alzheimer's disease, Lancet 348 (1996) 483. A.M. Saunders, C. Hulette, K.A. Welsh-Bohmer, D.E. Schmechel, B. Crain, J.R. Burke, M.J. Alberts, W.J. Strittmatter, J.C.S. Breitner, S.V. Scott, P.C. Gaskell Jr., M.A. Pericak-Vance, A.D. Roses, Speci¢city, sensitivity, and predictive value of apolipoprotein-E genotyping for sporadic Alzheimer's disease, Lancet 348 (1996) 90^93. K.A. Welsh-Bohmer, M. Gearing, A.M. Saunders, A.D. Roses, S. Mirra, Apolipoprotein E genotypes in a neuropathological series from the consortium to establish a registry for Alzheimer's disease, Ann. Neurol. 42 (1997) 319^325. M.R. Farlow, D.K. Lahiri, J. Poirier, J. Davignon, L. Schneider, S.L. Hui, Treatment outcome of tacrine therapy

BBADIS 61960 4-7-00

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

depends on apolipoprotein genotype and gender of the subjects with Alzheimer's disease, Am. Acad. Neurol. 50 (1998) 669^677. P. Seubert, D. Galasko, M. Boss, Use of CSF-based Markers in the diagnosis of Alzheimer's disease, Pharmacological treatment of Alzheimer's disease, Mol. Neurobiol. Found. 15 (1997) 345^365. K. Blennow, A. Wallin, O. Hager, Low frequency of postlumbar puncture headache in demented patients, Acta Neurol. Scand. 88 (1993) 221^223. M. Vandermeeren, M. Maercken, E. Vanmechelen, J. Six, A. Van de Voorde, J. Martin, P. Cras, J. Neurochem. 61 (1993) 1828^1834. D. Galasko, L. Chang, R. Motter, C.M. Clark, J. Kaye, D. Knopman, R. Thomas, D. Kholodenko, D. Schenk, I. Lieberburg, B. Miller, R. Green, R. Basherad, L. Kertiles, M.A. Boss, P. Seubert, High Cerebrospinal £uid tau and low amyloid b42 levels in the clinical diagnosis of Alzheimer disease and relation to apolipoprotein E genotype, Arch. Neurol. 55 (1998) 937^945. M. Kanai, E. Matsubara, K. Iso, K. Urakami, K. Nakashima, H. Arai, H. Sasaki, K. Abe, T. Iwatsubo, T. Kosaka, M. Watanabe, Y. Tomidokoro, M. Shizuka, K. Mizushima, T. Nakamura, Y. Igeta, Y. Ikeda, M. Amari, T. Kawarabayashi, K. Ishiguro, Y. Harigaya, K. Wakabayaashi, K. Okamoto, S. Hirai, M. Shoji, Longitudinal study of cerebrospinal £uid levels of tau, AL1-40, and AL1-42(43) in Alzheimer's disease: A study in Japan, Ann. Neurol. 44 (1998) 18^26. R. Motter, C. Vigo-Pelfrey, D. Kholodenko, R. Barbour, K. Johnson-Wood, D. Galasko, L. Chang, B. Miller, C. Clark, R. Green, D. Olson, P. Southwick, R. Wolfert, B. Munroe, I. Lieberburg, P. Seubert, D. Schenk, Reduction of LAmyloid42 in the cerebrospinal £uid of patients with Alzheimer's disease, Ann. Neurol. 38 (1995) 643^648. C. Vigo-Pelfrey, P. Seubert, R. Barbour, C. Blomquist, M. Lee, D. Lee, F. Coria, L. Chang, B. Miller, I. Lieberburg, D. Shenk, Elevation of microtubule-associated tau in the cerebrospinal £uid of patients with Alzheimer's disease, Neurology 45 (1995) 788^793. W.A. Munroe, P.C. Southwick, L. Change, D.W. Scharre, C.L. Echols, P.C. Fu, J.M. Whaley, R.L. Wolfert, Tau protein in cerebrospinal £uid as an aid in the diagnosis of Alzheimer's disease, Ann. Clin. Lab. Sci. 25 (1995) 207^217. H. Arai, M. Terajima, M. Miura, S. Higuchi, T. Muramatsu, N. Machida, H. Seiki, S. Takase, C.M. Clark, V.M.-Y. Lee, J.Q. Trojanowski, H. Sasaki, Tau in cerebrospinal £uid: A potential diagnostic marker in Alzheimer's disease, Ann. Neurol. 38 (1995) 649^652. D. Galasko, C. Clark, L. Chang, B. Miller, R.C. Green, R. Motter, P. Seubert, Assessment of CSF levels of tau protein in mildly demented patients with Alzheimer's disease, Neurology 48 (1997) 632^635. M. Riemenschneider, K. Buch, M. Schmolke, A. Kurz, W.G. Guder, Cerebrospinal protein tau is elevated in early Alzheimer's disease, Neurosci. Lett. 212 (1996) 209^211.

199

[33] R.E. Tato, A. Frank, A. Hernanz, Tau protein concentration in cerebrospinal £uid of patients with dementia of the Alzheimer's type, J. Neurol. Neurosurg. Psych. 59 (1995) 280^283. [34] M. Jensen, H. Basun, L. Lannfelt, Increased cerebrospinal tau in patients with Alzheimer's disease, Neurosci. lett. 186 (1995) 189^191. [35] M. Blomberg, M. Jensen, H. Basun, L. Lannfelt, L. OlofWahlund, Increasing cerebrospinal £uid tau levels in subgroup of Alzheimer's patients with Apolipoprotein E allele E4 during 14 months follow-up, Neurosci. Lett. 214 (1996) 163^166. [36] N. Rosler, I. Wichart, C. Bancher, K.A. Jellinger, Tau protein and apolipoprotein E in CSF diagnostics of Alzheimer's disease: impact on non Alzheimer's dementia?, J. Neural Transm. 47 (1996) 259^266. [37] C. Hock, S. Golombowski, W. Naser, F. Muller-Spahn, Increased levels of tau protein in cerebrospinal £uid of patients with Alzheimer's disease - correlation with degree of cognitive impairment, Ann. Neurol. 37 (1995) 414^415. [38] H. Mori, K. Hosoda, E. Matsubara, T. Nakamoto, Y. Furiya, R. Endoh, M. Usami, M. Shoji, S. Maruyama, S. Hirai, Tau in cerebrospinal £uids: establishment of the sandwich ELISA with antibody speci¢c to the repeat sequence in tau, Neurosci. Lett. 186 (1995) 181^183. [39] M. Riemenschneider, K. Buch, M. Schmolke, A. Kurz, W.G. Guder, Lancet 350 (1997) 784. [40] N. Rosler, I. Wichart, K.A. Jellinger, Total tau protein immunoreactivity in lumbar cerebrospinal £uid of patients with Alzheimer's disease, J. Neurol. Neurosurg. Psych. 60 (1996) 237^238. [41] K. Blennow, A. Wallin, H. Agren, C. Spenger, J. Siegried, E. Vanmechelen, Tau protein in cerebrospinal £uid a biochemical marker for axonal degeneration in Alzheimer's disease?, Mol. Chem. Neuropathol. 26 (1995) 231^245. [42] I. Skoog, E. Vanmechelen, L.A. Andreasson, B. Palmertz, P. Davidsson, C. Hesse, K. Blennow, A population-based study of tau protein and ubiquitin in cerebrospinal £uid in 85-yearolds: Relation to severity of dementia and cerebral atrophy, but not to the Apolipoprotein E4 allele, Neurodegeneration 4 (1995) 433^442. [43] K. Isoe, K. Urakami, T. Shimomura, Y. Wakutani, Y. Ji, Y. Adachi, K. Takahashi, Tau proteins in cerebrospinal £uid from patients with Alzheimer's disease: A longitudinal study, Dementia 7 (1996) 175^176. [44] H. Arai, S. Higuchi, H. Sasaki, Apolipoprotein E genotyping and cerebrospinal £uid tau protein : Implications for the clinical diagnosis of Alzheimer's disease, Gerontology 43 (1997) 2^10. [45] H. Vanderstichele, K. Blennow, N. D'Heuvaert, M. Buyse, P. Seubert, A. Van De Voorde, E. Vanmechelen, Development of a speci¢c diagnostic for measurement of b-Amyloid (1-42) in CSF, in: Program on AD and PD, 4th International Meeting, Eliat, Israel, 1997. [46] R.J. Ellis, P. Seubert, R. Motter, D. Galasko, R. Deutsch, R.K. Heaton, M.P. Heyes, J.A. McCutchan, J. Hampton

BBADIS 61960 4-7-00

200

M.A. Boss / Biochimica et Biophysica Acta 1502 (2000) 188^200

Atkinson, I. Grant, HIV Neurobehavioral Research Center Group, Cerebrospinal £uid tau protein is not elevated in HIV-associated neurologic disease in humans, Neurosci. Lett. 254 (1998) 1^4. [47] A. Tamaoka, N. Sawamura, T. Fukushima, S.I. Shoji, E. Matsubara, M. Shoji, S. Hirai, Y. Furiya, R. Endhoh, H. Mori, J. Neurol. Sci. 148 (1997) 41^45. [48] N. Ida, T. Hartmann, J. Pantel, J. Schroder, P. Zerfass, H. Forst, R. Sandbrink, C.L. Masters, K. Beyreuther, Analysis of heterogeneous Ba4 peptides in human cerebrospinal £uid and blood by a newly developed sensitive Western blot assay, J. Biol. Chem. 271 (1996) 22908^22914.

[49] P.H. St. George-Hyslop, Role of genetics in tests of genotype, status, and disease progression in early-onset Alzheimer's disease, Neurobiol. Aging 19 (1998) 133^137. [50] R. Sherrington, E.I. Rogaev, Y. Liang, E.A. Rogaeva, G. Levesque, M. Ikeda, H. Chi, C. Lin, G. Li, K. Holman, T. Tsuda, L. Mar, J.-F. Foncin, A.C. Bruni, M.P. Montesi, S. Sorbi, I. Rainero, L. Pinessi, L. Nee, I. Chumakov, D. Pollen, A. Brookes, P. Sanseau, R.J. Polinsky, W. Wasco, H.A.R. Da Silva, J.L. Haines, M.A. Pericak-Vance, R.E. Tanzi, A.D. Roses, P.E. Fraser, J.M. Rommens, P.H. St. George-Hyslop, Nature 375 (1995) 754^760.

BBADIS 61960 4-7-00