Summary of NIH Workshop on Dementia of Alzheimer’s Disease and Epilepsy: Converging Mechanisms

Hilton Washington DC/Rockville Executive Meeting Center
Rockville, Maryland
October 15-17, 2008

Introduction

Alzheimer's disease (AD) is a devastating neurological disease that now affects over 5 million individuals in the US and is predicted to affect over 12 million by the year 2050. The hallmarks of AD include progressive memory loss, behavioral alterations, accumulation of cerebral amyloid-ß (Aß ) deposits, and formation of neurofibrillary tangles (reviewed in ref. 1). These are accompanied by neurodegenerative changes, including synaptic loss and neuronal cell death, in the entorhinal cortex and hippocampus, brain regions key for learning and memory. Much recent AD research has focused on understanding mechanisms of A metabolism and neurotoxicity, and treatments aimed at reducing Aß accumulation are in clinical trials.^{2, 3} For now, though, the armamentarium of FDA approved AD therapeutics remains small, with only four drugs currently in widespread use (three cholinesterase inhibitors and the NMDA receptor agonist memantine). None of these drugs target the primary molecular pathology of the disorder, but instead are employed to augment cognitive performance, although they fail to substantially help most patients.

An alternative therapeutic approach is suggested by the fact that AD is associated with not only Aß deposition but also an increased incidence of unprovoked epileptic seizures. Until recently, the presence of seizures in patients with AD was not assigned any particular pathogenic significance, remained largely unexplained, and was generally believed to be coincidental. However, recent work in transgenic mice has shown that (1) high levels of A elicit epileptiform activity together with cognitive deficits, and (2) experimental manipulations that prevented this activity also prevented cognitive deficits. Notably, Aß -induced epileptiform activity and cognitive deficits occurred in the absence of cell loss. These observations raise the possibility that Aß -induced alterations in circuit activity contribute to cognitive deficits in human AD patients. Such changes might present a more reversible target for disease intervention than neuronal loss. Since a large number of drugs are already available that reduce or eliminate seizure activity in epilepsy patients, further exploration of the role of aberrant circuit activity in AD seems warranted.

Seizures in Human AD patients

A number of studies have shown that seizures are more common in AD patients than in the general population, even after correcting for the age-associated increase in seizure incidence that occurs in the latter.^4-13 For example, 7-21% of patients with sporadic AD are estimated to experience at least one unprovoked, clinically apparent seizure during their illness. Frank seizures are nonetheless relatively rare in sporadic AD patients. However, their incidence increases substantially in patients with early onset AD: the relative risk of unprovoked seizures reaches 87-fold for those whose dementia begins in their 50s, while in more elderly individuals, the risk declines but remains elevated by nearly three-fold compared with age-matched unaffected individuals. High seizure incidence is also seen in patients with Down's syndrome, most of whom bear an extra copy of the amyloid precursor protein (APP) gene and develop early-onset AD; 84% of Down's patients with dementia have clinical seizures.¹⁴ Increased epileptiform activity has also been reported in first-order relatives of patients with early onset AD¹⁵ and in non-demented apolipoprotein E4 (apoE4) carriers.¹⁶

AD patients and particularly patients with Lewy bodies can show remarkable fluctuations in functionality over the course of days or even hours.^17-19 Such fluctuations are unlikely to result from fleeting changes in the neuronal number in impaired brain networks, but could result from dynamic aberrations in physiological circuit activity, including actual seizures. Indeed, transient episodes of amnesia have been associated with non-convulsive epileptiform EEG discharges in AD patients, and both disturbances could be prevented by anti-epileptic drug (AED) treatment.¹⁹ Treatment with anti-epileptic drugs can improve cognitive function in epilepsy patients and reduce seizure frequency. Thus, it seems plausible that AED treatment could also improve cognitive function in AD patients, or at least in the subpopulation of them who experience clinical or sub-clinical seizures. However, many AEDs can impair cognitive functions (reviewed in ref. 20), making selection of drugs with an optimal side effect profile critical.

Seizures and Cognition in Mouse Models of AD

The hippocampus is among the earliest brain regions affected by Aß deposition and circuit dysfunction in AD. It is also an important site of seizure initiation. To explore the potential relationship between Aß, aberrant circuit activity, and cognition, Palop and colleagues²¹ studied human APP transgenic mice (line J20) with high levels of A in the brain. Previous studies had shown that these mice have AD-like cellular abnormalities, including amyloid plaques, and neuritic dystrophy, as well as cognitive deficits similar to those seen in human AD (e.g., deficits in spatial learning and memory). Neuropathological analysis of the hippocampal formation revealed extensive network remodeling at the anatomical and biochemical levels. EEG recordings revealed that these mice exhibit spontaneous non-convulsive seizures, which were accompanied by hippocampal hyperexcitability as measured by depth electrodes. Importantly, most of this abnormal circuit activity was not obvious by simple inspection of the mice (i.e., it was not accompanied by abnormal movements, such as tonic or clonic activity).

The presence of seizures and cognitive deficits in the mouse model may be linked. Some evidence for a converging mechanism has been obtained by genetic deletion experiments. Cognitive deficits in hAPP mice can be prevented by reducing their levels of tau, achieved by crossing the mice onto a Tau-deficient background.²² This manipulation blocked Aß-induced behavioral and neuronal deficits without affecting either APP or Aß. It also eliminated essentially all evidence of seizure activity and hippocampal cellular changes associated with that activity.

Several other groups have independently reported increased seizure susceptibility and/or epileptiform activity in other AD model transgenic mice, including lines carrying the Swedish mutation (APPsw) or both APPsw and a mutated human presenilin (PS1) gene.^23-25 In addition, microelectrode studies of brain slices from the cortex of the APPsw/mutant PS1 mouse model showed a large percentage of abnormally hyperactive neurons immediately surrounding amyloid plaques.²⁶

In addition to the evidence that the pathological accumulation of Aß can lead to an epileptogenic lesion, a recent study of normal hippocampus has revealed that the Aß peptide can in turn be released from presynaptic terminals by repetitive stimulation and kainic-acid-induced seizure activity.²⁷ This amyloidogenic effect suggests that seizures in the AD brain could be especially deleterious throughout the entire course of the disease.

Shared Phenotypes and Mechanisms in AD and Epilepsy

Overlapping Clinical Features in AD and Temporal Lobe Epilepsy. AD and temporal lobe epilepsy (TLE) have a number of shared clinical features and risk factors. As is true for AD patients, TLE patients have memory deficits that are accompanied by cell loss and other structural rearrangements in the hippocampus.²⁸ Both AD and TLE can show fluctuations in cognitive and behavioral functions, although the extent of overlap between the underlying mechanisms remains to be determined. There are shared predictors of cognitive decline in the two diseases, including smaller hippocampal volume and lower baseline IQ.²⁹ In addition, TLE patients carrying the apoE4 allele show a correlation between duration of epilepsy and poorer memory performance, a relationship not observed in non apoE4 carriers.^{30, 31}

While patients with epilepsy have been shown to exhibit many risk factors for cognitive impairment identified in the aging/dementia literature,²⁹ these factors have rarely been examined in relationship to cognition in the epilepsy patient population. Vascular risk factors are also elevated in epilepsy patients, but again have rarely been investigated in relation to cognitive status in epilepsy.

Cellular and Molecular Mechanisms. Functional MRI studies in humans show that AD is accompanied by changes in circuit activity, including hyperactivation of the hippocampus and reduced functional connectivity between hippocampus and neocortex.^{32, 33} However, analyses of cellular architecture in pathological specimens from human AD and TLE patients indicate that the specific elements of circuitry reorganization are highly variable in both groups. Thus, while certain individual cases of AD exhibit pathology similar to that seen in certain cases of TLE, overall the patterns of degeneration seem rather different in the two diseases, at least on first sight. For example, although there is CA1 degeneration in the hippocampus of both AD and TLE patients, the granule cell layer is typically well preserved in AD but greatly diminished in TLE. However, although granule cells resist degeneration in AD until late stages, their molecular profile suggests functional abnormalities even at early stages of the disease.³⁴

Similarly, the cellular changes observed in AD and convulsant-induced epilepsy mouse models are overlapping but not identical. Changes seen in the hippocampi of hAPP mice include NPY/GABAergic neuronal sprouting, enhanced synaptic inhibition, and down-regulation of calbindin.^{21, 34} Some of these changes are also present in humans with AD and can be recapitulated in nontransgenic mice by challenge with the convulsant drug kainic acid.²¹ It is currently unclear which of the changes found in AD models actually represent compensatory reactions to seizure activity. hAPP mice, like epilepsy model mouse models, show evidence of ectopic mossy fiber sprouting onto interneurons within the granule cell layer. However, neither hAPP nor APPsw³⁵ mutant mice show the intense mossy fiber sprouting into the inner molecular layer of the dentate gyrus or loss of somatostatin neurons seen in epilepsy models.

Clinical Trial Design

The findings discussed above raise the possibility that AEDs could be beneficial in AD patients, or at least in the subpopulation of them that experience clinical or subclinical seizures. If non-convulsive epileptiform seizures contribute to confusional spells of wandering in advanced stages of the disease, AED's might be effective in minimizing this episodic behavior. If seizure activity induces maladaptive plasticity in hippocampal circuitry important for learning and memory, early AED treatment might be an important modality in slowing disease progression. The first step toward planning a clinical trial of AEDs in AD would be to address the following issues:

Additional data to be gathered before beginning pilot trials. It will be critical to choose an animal model, such as the hAPP mouse, in which there is clear evidence that seizure-like EEG events are associated with impaired cognition and behavioral alterations, and demonstrate that a specific AED both reduces or terminates these events and improves cognition. In addition, it would be ideal, although perhaps not essential, to demonstrate that at least a subpopulation of AD patients also have seizure-like events associated with a clinically measurable change in behavior. This might be accomplished by prolonged ambulatory EEG monitoring in a pilot cohort of patients with AD, in particular those experiencing episodic confusional spells. Finally, it will be necessary to identify a clinical outcome measure for these seizure-like events and/or their behavioral sequelae.

Trial duration. Both short- and long-term trials have been used to test drug efficacy in AD. Short-term (e.g. 12-24 week) trials look for improvement of cognitive or behavioral symptoms, whereas longer-term (e.g. 18 month) trials look for disease modification, such as slowing of the course of cognitive decline. To test AEDs in AD, short-term trials might assess changes in the frequency of electrical events or associated symptoms, as such changes might be expected to occur in a relatively short period. However, this trial design is complicated by the fluctuating nature of electrical events and symptoms. Long-term trials might measure more global outcomes (e.g. through a mental status exam) before and after treatment or at intervals during treatment. But because long-term trials are complicated and expensive, the case for their initiation would be strengthened by more compelling evidence of aberrant electrical activity in the study population.

Target pathology to measure. The most hypothesis-relevant target pathology to monitor would be sub-clinical epileptiform activity. While the easiest and least invasive way to detect such activity is by traditional scalp EEG, this method fails to detect seizure events occurring in deep brain structures such as the mesial temporal lobe. Greater sensitivity can be obtained using recently developed computational and statistical analytic methods and/or by simultaneous EEG and functional magnetic resonance imaging (fMRI), although this cannot be done on an ambulatory basis, and thus depends upon fortuitous capture of a seizure-like event during a brief monitoring session. Alternatively, seizure activity in deep temporal lobe pathways can be better detected by implantation of depth or foramen ovale EEG electrodes, but it may be difficult to obtain approval and consent for such invasive methods in a cognitively impaired population.

With regard to cognitive and behavioral outcome measures, one of the simplest to collect is a clinician's interview-based impression of cognitive status along with caregiver input. Additional measures might include a mental status exam, severe impairment battery, or disability battery. Alternatively, there are event-based measures that have been associated with non-convulsive seizures in epilepsy patients, such as episodes of momentary confusion, emotional lability, and delusions. These may be more readily associated with aberrant electrical activity in patients with known epilepsy than in those without. Epileptiform activity-associated cognitive fluctuations might also be detected using sustained concentration and attention tasks, tests of vigilance, delayed recall, or even a driving test with an intrinsic learning component.

Choice of patient population to study. If excess A causes aberrations in circuit activity, one might speculate that such aberrations will occur in all AD patients regardless of molecular etiology. However, AD is clearly a multifactorial disease, and numerous genetic loci are known to affect the susceptibility to epileptic seizures. For example, the genetic background of inbred strains of mice has a significant effect on the ability of convulsants to elicit seizures. Hence, it may be of interest to first screen patients with AD of different etiologies for non-convulsive seizure activity in order to identify particularly susceptible individuals showing EEG evidence of epileptiform abnormalities, and then use this group for clinical trials.

Particularly susceptible subpopulations may be stratified by medical history, including patients with early-onset AD, patients with sporadic AD who have had one or more seizures, and AD patients with prominent fluctuations in cognition or behavior. Patients with early-onset AD offer a clear advantage in terms of relative homogeneity of disease mechanism. These patients would be younger and healthier, have fewer co-morbidities, and may be generally more eager to participate in clinical trials. In addition, onecould study these patients early in the course of their disease when it might be easier to identify seizure-associated cognitive fluctuations. The disadvantage of this subpopulation is that these patients are relatively rare and thus more costly to recruit and bring to study centers. In addition, the mechanism of disease in these patients may be different than in sporadic AD.

Even if one decides to study the rare familial AD population, heterogeneity within this population must still be borne in mind. AD can result from mutations in at least three genes, APP, presenilin 1 or presenilin 2 (PS1 and PS2). Both the specific mechanisms and clinical phenotype of the disease process may differ somewhat depending not only on the gene mutated, but also the type of mutation in a given gene. Finally, in both familial and sporadic AD, the potential influence of apoE genotype on the mechanism and course of disease must be considered (with the apoE4/E4 allele combination giving the youngest age of onset, for example).

Choice of AEDs to test first. There are nearly a dozen possible AEDs available for a clinical trial, and based on the epilepsy population, in particular those with TLE, many of the drugs are effective in only a small percentage of patients. Indeed, as a whole, up to 50% of patients with epilepsy suffer from medically intractable seizures. An AED that is effective in one type of seizure or individual may exacerbate seizures in another. Thus it is predictable that seizures in AD, if they are common, may respond in a very heterogeneous manner, or not at all, to any of the currently approved AEDs. AEDs are particularly problematic in treating the elderly, with multiple relevant side effects such as cognitive disturbances and drug interactions. While there have been anecdotal reports of AD patients being treated with various AEDs, there is little indication of which drug or class of drugs might be the most favorable in any given patient. One recommendation is to identify an effective AED in the animal model and to use this evidence in assisting the prioritization and selection of AEDs for clinical study.

Recommendations for Moving Forward

While there is no doubt that AD patients have signaling defects within critical neural networks, the precise nature of this dysfunction remains to be defined. The following issues need to be addressed in AD patients and related animal models:

• Determine if AD of different etiology results in the same type of network dysfunction, and whether it includes epileptiform activity.
• Identify when, where, and how the network dysfunction emerges in different animal models and human forms of AD.
• Understand which of the molecular, cellular, and circuit changes seen in mouse and human AD are primary to the disease process and which are compensatory.
• Investigate changes in glial, microglial, and vascular cells that may accompany changes in neurons.
• Identify a reliable and non-invasive outcome measure for assessing network dysfunction in human AD.
• Ask if network dysfunction responds to any of the available or investigational anticonvulsants.
• Determine when treatment has to be initiated to prevent progressive deterioration of cognitive function.
• Learn if normalization of neural network dysfunction prevents, stalls, or reverses the neurodegenerative process.

REFERENCES

1. Hardy J. A hundred years of Alzheimer's disease research. Neuron. 2006;52:3-13

2. Brody DL, Holtzman DM. Active and passive immunotherapy for neurodegenerative disorders. Annu Rev Neurosci. 2008;31:175-193

3. Wisniewski T, Konietzko U. Amyloid-beta immunisation for Alzheimer's disease. Lancet Neurol. 2008;7:805-811

4. Amatniek JC, Hauser WA, DelCastillo-Castaneda C, Jacobs DM, Marder K, Bell K, Albert M, Brandt J, Stern Y. Incidence and predictors of seizures in patients with Alzheimer's disease. Epilepsia. 2006;47:867-872

5. Hauser WA, Morris ML, Heston LL, Anderson VE. Seizures and myoclonus in patients with Alzheimer's disease. Neurology. 1986;36:1226-1230

6. Romanelli MF, Morris JC, Ashkin K, Coben LA. Advanced Alzheimer's disease is a risk factor for late-onset seizures. Arch Neurol. 1990;47:847-850

7. Mendez M, Lim G. Seizures in elderly patients with dementia: Epidemiology and management. Drugs Aging. 2003;20:791-803

8. Risse SC, Lampe TH, Bird TD, Nochlin D, Sumi SM, Keenan T, Cubberley L, Peskind E, Raskind MA. Myoclonus, seizures, and paratonia in Alzheimer disease. Alzheimer Dis Assoc Disord. 1990;4:217-225

9. Lozsadi DA, Larner AJ. Prevalence and causes of seizures at the time of diagnosis of probable Alzheimer's disease. Dement Geriatr Cogn Disord. 2006;22:121-124

10. Mendez MF, Catanzaro P, Doss RC, R AR, Frey WH, 2nd. Seizures in Alzheimer's disease: Clinicopathologic study. J Geriatr Psychiatry Neurol. 1994;7:230-233

11. Volicer L, Smith S, Volicer BJ. Effect of seizures on progression of dementia of the Alzheimer type. Dementia. 1995;6:258-263

12. Larner AJ, Doran M. Clinical phenotypic heterogeneity of Alzheimer's disease associated with mutations of the presenilin-1 gene. J Neurol. 2006;253:139-158

13. Snider BJ, Norton J, Coats MA, Chakraverty S, Hou CE, Jervis R, Lendon CL, Goate AM, McKeel DW, Jr., Morris JC. Novel presenilin 1 mutation (s170f) causing Alzheimer disease with Lewy bodies in the third decade of life. Arch Neurol. 2005;62:1821-1830

14. Lai F, Williams RS. A prospective study of Alzheimer disease in Down syndrome. Arch Neurol. 1989;46:849-853

15. Ponomareva NV, Selesneva ND, Jarikov GA. EEG alterations in subjects at high familial risk for Alzheimer's disease. Neuropsychobiology. 2003;48:152-159

16. Ponomareva NV, Korovaitseva GI, Rogaev EI. EEG alterations in non-demented individuals related to apolipoprotein e genotype and to risk of Alzheimer disease. Neurobiol Aging. 2008;29:819-827

17. Bradshaw J, Saling M, Hopwood M, Anderson V, Brodtmann A. Fluctuating cognition in dementia with Lewy bodies and Alzheimer's disease is qualitatively distinct. J Neurol Neurosurg Psychiatry. 2004;75:382-387

18. Bachman D, Rabins P. "Sundowning" and other temporally associated agitation states in dementia patients. Annu Rev Med. 2006;57:499-511

19. Rabinowicz AL, Starkstein SE, Leiguarda RC, Coleman AE. Transient epileptic amnesia in dementia: A treatable unrecognized cause of episodic amnestic wandering. Alzheimer Dis Assoc Disord. 2000;14:231-233

20. Loring DW, Marino S, Meador KJ. Neuropsychological and behavioral effects of antiepilepsy drugs. Neuropsychol Rev. 2007;17:413-425

21. Palop JJ, Chin J, Roberson ED, Wang J, Thwin MT, Bien-Ly N, Yoo J, Ho KO, Yu GQ, Kreitzer A, Finkbeiner S, Noebels JL, Mucke L. Aberrant excitatory neuronal activity and compensatory remodeling of inhibitory hippocampal circuits in mouse models of Alzheimer's disease. Neuron. 2007;55:697-711

22. Roberson ED, Scearce-Levie K, Palop JJ, Yan F, Cheng IH, Wu T, Gerstein H, Yu GQ, Mucke L. Reducing endogenous tau ameliorates amyloid beta-induced deficits in an Alzheimer's disease mouse model. Science. 2007;316:750-754

23. Westmark CJ, Westmark PR, Beard AM, Hildebrandt SM, Malter JS. Seizure susceptibility and mortality in mice that over-express amyloid precursor protein. Int J Clin Exp Pathol. 2008;1:157-168

24. Moechars D, Lorent K, Van Leuven F. Premature death in transgenic mice that overexpress a mutant amyloid precursor protein is preceded by severe neurodegeneration and apoptosis. Neuroscience. 1999;91:819-830.

25. Moechars D, Lorent K, De Strooper B, Dewachter I, Van Leuven F. Expression in brain of amyloid precursor protein mutated in the alpha-secretase site causes disturbed behavior, neuronal degeneration and premature death in transgenic mice. EMBO J. 1996 6:1265-74.

26. Busche MA, Eichhoff G, Adelsberger H, Abramowski D, Wiederhold KH, Haass C, Staufenbiel M, Konnerth A, Garaschuk O. Clusters of hyperactive neurons near amyloid plaques in a mouse model of Alzheimer's disease. Science. 2008;321:1686-1689

27. Cirrito JR, Yamada KA, Finn MB, Sloviter RS, Bales KR, May PC, Schoepp DD, Paul SM, Mennerick S, Holtzman DM. Synaptic activity regulates interstitial fluid amyloid-beta levels in vivo. Neuron. 2005;48:913-922

28. Hoppe C, Elger CE, Helmstaedter C. Long-term memory impairment in patients with focal epilepsy. Epilepsia. 2007;48 Suppl 9:26-29

29. Hermann B, Seidenberg M, Lee EJ, Chan F, Rutecki P. Cognitive phenotypes in temporal lobe epilepsy. J Int Neuropsychol Soc. 2007;13:12-20

30. Busch RM, Lineweaver TT, Naugle RI, Kim KH, Gong Y, Tilelli CQ, Prayson RA, Bingaman W, Najm IM, Diaz-Arrastia R. Apoe-epsilon4 is associated with reduced memory in long-standing intractable temporal lobe epilepsy. Neurology. 2007;68:409-414

31. Gambardella A, Aguglia U, Chifari R, Labate A, Manna I, Serra P, Romeo N, Sibilia G, Lepiane E, Russa AL, Ventura P, Cittadella R, Sasanelli F, Colosimo E, Leggio U, Zappia M, Quattrone A. Apoe epsilon4 allele and disease duration affect verbal learning in mild temporal lobe epilepsy. Epilepsia. 2005;46:110-117

32. Celone KA, Calhoun VD, Dickerson BC, Atri A, Chua EF, Miller SL, DePeau K, Rentz DM, Selkoe DJ, Blacker D, Albert MS, Sperling RA. Alterations in memory networks in mild cognitive impairment and Alzheimer's disease: An independent component analysis. J Neurosci. 2006;26:10222-10231

33. Liu Y, Wang K, Yu C, He Y, Zhou Y, Liang M, Wang L, Jiang T. Regional homogeneity, functional connectivity and imaging markers of Alzheimer's disease: A review of resting-state fmri studies. Neuropsychologia. 2008;46:1648-1656

34. Palop JJ, Jones B, Kekonius L, Chin J, Yu GQ, Raber J, Masliah E, Mucke L. Neuronal depletion of calcium-dependent proteins in the dentate gyrus is tightly linked to Alzheimer's disease-related cognitive deficits. Proc Natl Acad Sci U S A. 2003;100:9572-9577

35. Tanila HJ, Minkeviciene R, M. B. Dobszay MB, Zilberter M, Rheims S, Hartikainen J, Harkany T, Zilberter Y, Pitkänen A. Fibrillar ?-amyloid associates with epileptic seizures in APPswe/PS1de9 mice and changes excitability in cortical and hippocampal neurons. Soc Neurosci Abstr 2008;34:610.