Series: Seminal Neuroscience Papers 1978–2017
Science & Society
In Search of an Identity for Amyloid Plaques Tien-Phat V. Huynh1,2 and David M. Holtzman2,* In 1984, biochemists George Glenner and Caine Wong, in search of ‘a unique amyloid fibril precursor protein in the serum’ of Alzheimer disease (AD) patients, successfully isolated and sequenced the first 24 amino acids of a ‘cerebrovascular amyloid fibril protein b’ that we now know as the amyloid-b (Ab) peptide. This landmark paper laid the foundation for extensive research in the following decades that ultimately established the role of b-amyloidosis as a player in the pathogenesis of AD. Efforts to Purify the ‘b Protein’ From as early as the late 1800s, the presence of amyloid plaques (then described as ‘miliary foci’) was documented in the brain of elderly patients suffering from dementia. In 1906 (Figure 1), Alois Alzheimer reported the presence of a ‘peculiar substance’ in the brain of the late Auguste Deter, but the identity of the substance was to remain a mystery for nearly eight decades. In the late 1960s, the pathological accumulation of proteinaceous deposits of b-sheet-containing (amyloid) fibrils was described in the context of various clinical conditions, including systemic forms of amyloidosis (e.g., AL amyloid), Down syndrome (DS), and AD. Though researchers suspected some of these conditions might share certain pathologic entities, the limitations of biochemical techniques at the time (specifically those required to
produce a homogeneous protein preparation) prevented the identification of the protein species involved. In 1983, Allsop and colleagues described a method for isolating dense core neuritic plaques from post-mortem AD brains, and found the amino acid composition of the isolated species to be unique compared to any previously described amyloid protein [1]. However, possible contaminants in the preparation (as noted by the authors) impeded a definitive identification of the amino acid composition, and, in addition, no amino acid sequence was provided. Long-time researchers in the field of amyloidosis, George Glenner and Caine Wong, applied a slightly different approach. By focusing on cerebrovascular amyloidosis, they made the observation that the latter was ‘seen only in Alzheimer’s disease [and] adult Down’s syndrome individuals’. Thus, the authors hypothesized that the isolation and identification of ‘the cerebrovascular amyloid fibril protein in Alzheimer’s disease’ would lead to the discovery of a unique fibril precursor protein in the serum of these patients, which in turn could lead to a specific diagnostic serum test for AD. While their suspicion on the serum origin of the precursor protein turned out to be incorrect, Glenner and Wong were successful in their quest to identify the precursor protein (Ab), sequencing its first Nterminal 24 amino acids, as reported in their seminal 1984 paper in Biochemical and Biophysical Research Communications [2].
The Breakthrough Convinced of the vascular nature of the so-called amyloid precursor, the authors decided to enrich for amyloid-bearing meningeal vessels by dissecting out the meninges of autopsy-confirmed AD brains (six samples) and those from agematched controls (three samples). The AD brains were selected for their extensive cerebrovascular amyloidosis based on histological staining for amyloid.
Following sequential homogenization steps in sodium chloride and Tris-hydrochloride (0.05 [70_TD$IF]M), the amyloid-enriched material (monitored by polarization microscopy after Congo red staining) was further purified through treatment with collagenase, which removed a primary contaminant. Subsequent solubilization in the chaotropic salt guanidine hydrochloride (6 M), followed by chromatographic enrichment for the amyloid subunit, yielded a 4.2-kDa band on SDSurea PAGE gels. Interestingly, HPLC analysis identified two peaks, which were found to have identical amino acid sequence up to residue 24 (most likely corresponding to Ab1-40 and Ab1-42) based on amino-terminal sequencing. Of note, their report of a glutamine (instead of a glutamate) at position 11 was later corrected in their subsequent sequencing of amyloid isolated from DS meningio-vasculature (see below). Upon confirming the uniqueness of the peptide, Glenner et al. concluded that the novel protein ‘appears to be a biologic marker for the cerebrovascular amyloid fibril component of Alzheimer’s disease’. This latter prediction held true, for the most part, as low levels of Ab1-42 levels in the cerebrospinal fluid (CSF), reflective of sequestration into amyloid fibrils in the brain, now serve as a biomarker for the presence of amyloid plaque presence in the brain.
Additional Insights from Down Syndrome In the context of this discussion, it is important to mention another paper from Glenner and Wong that was published in the same journal just 3 months after their initial report of the ‘b protein’ from AD blood vessels [3]. Pathologic similarities between adult DS and AD brains had been observed as early as the 1920s, and George Glenner discussed his suspicion for a ‘chemical connection’ as early as 1979 in an article in Medical Hypotheses [4]. However, establishing
Trends in Neurosciences, August 2018, Vol. 41, No. 8
483
1906
Alois Alzheimer (1906) • DescripƟon
of the first AD paƟent Histopathologic descripƟon of amyloid plaques and neurofibriliary tangles
•
Glenner and Wong (1984) [2]
1984
• Isolation
of Aβ from meningeal vessels of AD patients • Amino acid sequence of Aβ (N-terminus to residue 24)
Glenner and Wong (1984) [3]
1985
• Isolation
of Aβ from meningeal vessels of patients with down syndrome • ConfirmaƟon of Aβ amino acid sequence (N-terminus to residue 24)
Masters et al. (1985) [5] • Isolation
1986
of amyloid plaque cores from post-mortem AD brains • CharacterizaƟon of Aβ on SDS-PAGE and HPLC, N-terminal squence
Kang et al. (1987) [6] • Mapping
1987
of Aβ sequence to APP locus, localizaƟon to chromosome 21 • Cloning of APP gene • Full amino acid sequence of APP
Tanzi et al. (1987) [7] • ParƟal
cloning of Aβ-containing locus • Mapping of Aβ sequence to chromosome 21
1991
Goate et al. (1991) [8] • IdenƟficaƟon
of the first APP mutaƟon (V717I) in an English kindred with early-onset AD (known as the London mutaƟon)
1992
Hardy and Higgins (1992) [9] • Proposal
of the amyloid cascade hypothesis
such connections was not possible until the key pathologic entity was defined. Using a similar approach to their prior work, Glenner et al. successfully isolated another ‘b protein’, this time from the meningio-vasculature of two adult DS patients. Strikingly, the identical amino acid sequence between this peptide and those isolated from AD patients confirmed Glenner’s prior suspicion about a connection between these two seemingly distinct entities. These results were confirmed and extended just a year later by Masters, [71_TD$IF]Beyreuther, and colleagues who employed a variety of different techniques to isolate dense core plaques from the brain of AD and DS patients [5]. Importantly, upon these findings, Glenner made several predictions, many of which have since proved remarkably prescient. Glenner and Wong noted, for instance: ‘It suggests that Down’s syndrome may be a predictable model for Alzheimer’s disease. Assuming the beta protein is a human gene product, it also suggests that the genetic defect in Alzheimer’s disease is localized on chromosome 21’ [3]. Among their many ramifications, the findings and predictions by Glenner and colleagues have been considered by many as foundational in leading eventually to the amyloid cascade hypothesis.
The Amyloid Cascade Hypothesis, circa 1992
In the ensuing race to clone the gene encoding the Ab sequence, Ab was found to be, in fact, the cleavage prodSherrington et al. (1995) [11] uct of a much larger precursor protein 1995 • GeneƟc mapping of PSEN1 locus on (b-amyloid precursor protein – APP), chromosome 14 which was, indeed, localized to chromo• IdenƟficaƟon of missense mutaƟons in some families with familial AD some 21 [6,7]. Genetic mutations in the [6_TD$IF]APP locus (near the Ab sequence) were subsequently found in some families Figure 1. [68_TD$IF]A Timeline of Key Discoveries in the Quest to Identify the Primary Component of Amyloid with an autosomal-dominant pattern of Plaques and Its Genetic Links to Early-Onset Alzheimer Disease. The numbers in the brackets refer to AD incidence [8]. The accumulating evithe corresponding citation associated with the findings. The primary paper in discussion by Glenner and Wong dence (Figure 1) led to the formal is highlighted in orange. See also [2,3,5–9,11]. 484
Trends in Neurosciences, August 2018, Vol. 41, No. 8
proposal of the amyloid cascade hypothesis (or the amyloid hypothesis) by John Hardy and Gerald Higgins in 1992. Hardy and Higgins postulated, ‘[The] deposition of amyloid b protein (AbP), the main component of the plaques, is the causative agent of Alzheimer’s pathology and the neurofibrillary tangles, cell loss, vascular damage, and dementia follow as a direct result of this deposition’ [9]. Around the same time, it became apparent to researchers that APP mutations were not found in all families, suggesting a genetically heterogeneous nature of dominantly inherited AD [10] and, by extension, pointing at possible heterogeneity in the underlying causes. Apart from the APP mutations, some ADlinked mutations were identified in genes outside chromosome 21, for instance those in the PSEN1 and PSEN2 loci on chromosome 14, which on the surface may seem to question the idea of an Ab-centric component of the disease. But in fact, the PSEN1 and PSEN2 loci are components of the proteolytic complex that cleaves APP, and mutations in them were later shown to alter APP processing and enhance the production of Ab42; therefore altogether, these mutations do seem to provide further support for (or at least be consistent with) an Ab-centric cause of AD [11].
The Amyloid Cascade Hypothesis, circa 2018 As noted by Hardy later on, the amyloid hypothesis ‘was intended to generate ideas and act as a framework for a research agenda, not to be a definitive statement’ [12]. In line with this idea, the amyloid hypothesis had been at the forefront of AD research efforts for the past 26 years, where parts of it were validated, and others revised or supplemented [13]. While most in the field believe that Ab appears necessary but not sufficient for the ultimate
development of dementia due to AD, some researchers have also expressed more basic reservations on the causative role of Ab in the disease’s pathology, or on whether the amyloid cascade hypothesis provides a useful overarching conceptualization of the pathophysiology. Doubts around the overall validity of the hypothesis have been argued by some to be validated by the initial failure of some therapeutic compounds that target APP metabolism or Ab itself in clinical trials in dementia due to AD. While the details of pathways driving the disease are still being worked on, the failure of some of the clinical trials targeting Ab/APP does not necessarily conflict with a causal component of Ab accumulation. One crucial aspect to consider is the timing of treatment and the progression profile of the disease. In fact, Ab accumulation in the brains of AD patients often precedes clinical symptoms by several decades, during a period known as ‘preclinical’ AD [14]. By the time people are diagnosed with dementia due to AD, Ab accumulation in the brain has already been evolving for an extended period, often around 20 years or so, and the tau phase of the disease is already taking over. With that in mind, it is conceivable that Ab accumulation is indeed the initiator (or one of the initiators) of AD pathogenesis, but by the time an AD patient becomes symptomatic, some irreversible neuronal loss has already occurred, rendering amyloid-targeting therapies less likely to be effective when given at the symptomatic stage of disease. Would targeting amyloid during the preclinical phase prevent or mitigate AD? An ongoing international study (Dominantly Inherited Alzheimer Network – DIAN) is designed to answer this question, at least in some contexts, by following families that carry known mutations that cause AD [15]. In addition to collecting longitudinal data on key biomarkers (i.e., Ab, tau), the
patients are also treated with amyloidtargeting compounds in hope of delaying (or preventing) the development of AD. This study is not only critical in assessing the validity of the amyloid hypothesis, but will also provide invaluable information on the temporal progression of the disease, which, regardless of the specific outcome, will help constrain models of the disease’s progression and advance clarifying the underlying causes. Ultimately, this progress in understanding the disease will hopefully provide concrete guidance for future interventions or preventative measures. Disclaimer Statement D.M.H. cofounded and is on the scientific advisory board of C2N Diagnostics. D.M.H. is on the scientific advisory board of Denali, Genentech, and Proclara. D. M.H. consults for AbbVie and Eli Lilly.
1 Medical Scientist Training Program (MSTP), Washington University School of Medicine, St. Louis, MO 63110, USA 2 Department of Neurology, Hope Center for Neurological
Disorders, Knight Alzheimer’s Disease Research Center, Washington University School of Medicine, St. Louis, MO 63110, USA *Correspondence:
[email protected] (D.M. Holtzman). https://doi.org/10.1016/j.tins.2018.06.002 References 1. Allsop, D. et al. (1983) The isolation and amino acid composition of senile plaque core protein. Brain Res. 259, 348– 352 2. Glenner, G.G. and Wong, C.W. (1984) Alzheimer’s disease: initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem. Biophys. Res. Commun. 120, 885–890 3. Glenner, G.G. and Wong, C.W. (1984) Alzheimer’s disease and Down’s syndrome: sharing of a unique cerebrovascular amyloid fibril protein. Biochem. Biophys. Res. Commun. 122, 1131–1135 4. Glenner, G.G. (1979) Congophilic microangiopathy in the pathogenesis of Alzheimer’s syndrome (presenile dementia). Med. Hypotheses 5, 1231–1236 5. Masters, C.L. et al. (1985) Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc. Natl. Acad. Sci. U. S. A. 82, 4245–4249 6. Kang, J. et al. (1987) The precursor of Alzheimer’s disease amyloid A4 protein resembles a cell-surface receptor. Nature 325, 733–736 7. Tanzi, R.E. et al. (1987) Amyloid beta protein gene: cDNA, mRNA distribution, and genetic linkage near the Alzheimer locus. Science 235, 880–884
Trends in Neurosciences, August 2018, Vol. 41, No. 8
485
8. Goate, A. et al. (1991) Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer’s disease. Nature 349, 704–706 9. Hardy, J.A. and Higgins, G.A. (1992) Alzheimer’s disease: the amyloid cascade hypothesis. Science 256, 184–185 10. St George-Hyslop, P.H. et al. (1990) Genetic linkage studies suggest that Alzheimer’s disease is not a single homogeneous disorder. Nature 347, 194–197 11. Sherrington, R. et al. (1995) Cloning of a gene bearing missense mutations in early-onset familial Alzheimer’s disease. Nature 375, 754–760 12. Hardy, J. (2006) Alzheimer’s disease: the amyloid cascade hypothesis: an update and reappraisal. J. Alzheimers Dis. 9, 151–153 13. Musiek, E.S. and Holtzman, D.M. (2015) Three dimensions of the amyloid hypothesis: time, space and ‘wingmen’. Nat. Neurosci. 18, 800–806 14. Jack, C.R., Jr and Holtzman, D.M. (2013) Biomarker modeling of Alzheimer’s disease. Neuron 80, 1347– 1358 15. Bateman, R.J. et al. (2017) The DIAN-TU Next Generation Alzheimer’s prevention trial: Adaptive design and disease progression model. Alzheimers Dement. 13, 8–19
Series: Seminal Neuroscience Papers 1978–2017
Science & Society
Synaptic Receptor Diversity Revealed Across Space and Time Sharon A. Swanger1 and Stephen F. Traynelis2,* In their 1994 paper, Monyer et al. described unique functional properties and cell type-specific expression of NMDA receptor (NMDAR) GluN2 subunits, suggesting that the roles of NMDAR vary across brain regions and throughout development. This influential work not only provided insight into how molecular diversity impacts synapse function, but also continues to inform new 486
Trends in Neurosciences, August 2018, Vol. 41, No. 8
approaches for modulating brain that modify the molecular composition of synapses during learning. Cloning of circuits. The capacity for learning and memory in mammals is one of the most striking adaptations that has occurred during vertebrate evolution. A massive expansion of the synaptic proteome created almost boundless synapse diversity, enabling neuron ensembles to have highly specialized functions, thus allowing for the complex information-processing capabilities underlying higher cognition. Glutamate is the principle excitatory neurotransmitter in the mammalian brain, and NMDARs are a family of ionotropic glutamate receptors that mediate excitatory synaptic transmission. NMDARs are nonselective cation channels with unique biophysical properties that, among other roles, promote long-term synaptic plasticity (the cellular substrate of learning and memory). A 1994 paper by Monyer et al. [1] made timely discoveries demonstrating how diversity within the NMDAR family contributes to synapse function through the distinctive expression patterns and functional properties of four NMDAR subtypes. NMDARs are multifaceted signaling complexes that allow transmembrane cation influx in response to neurotransmitters (glutamate and glycine) as well as to depolarization, which reduces blocking of the channel pore by extracellular Mg2 + [79_TD$IF]. This combination of properties enables NMDARs to detect coincident presynaptic and postsynaptic neuron depolarization, a key mechanism underlying synaptic plasticity. In addition, NMDARs control the time course of synaptic transmission due to their slow deactivation kinetics [2], which expands the time window for detecting multiple synaptic events. Ca2+ influx through NMDARs is essential for the activity-induced posttranslational modifications, cytoskeletal rearrangements, and gene expression
five NMDAR genes during the early 1990s led to discoveries of the GluN1 and GluN2A-2D subunits as well as the requirement of both GluN1 and GluN2 subunits for forming functional NMDARs [3–7]. Subsequent studies determined that NMDARs are heterotetrameric assemblies of two GluN1 subunits and two GluN2 subunits of any type [8,9]. From an evolutionary perspective, the four genes encoding GluN2A–2D formed and diverged early during vertebrate evolution, and were then maintained for >500 million years [10], suggesting there is an advantage for organisms to have four different GluN2 subtypes as opposed to just one. The 1994 Monyer et al. paper from Peter Seeburg’s lab provided two key advances for understanding NMDAR diversity: the visualization of GluN2 mRNA at the cellular level in brain tissue, and detailed functional analyses of all four GluN2 subunits. A prior study showed spatial and temporal differences in GluN2A–2D expression in rodents; GluN2B and GluN2D mRNA are expressed embryonically, whereas GluN2A and GluN2C expression begins after birth [11]. In adulthood, GluN2A and GluN2B remain widely expressed, whereas GluN2C and GluN2D expression is more restricted. Monyer et al. used high-resolution emulsion autoradiography to demonstrate striking cell type-specific expression of GluN2A-2D mRNA within particular brain regions. In the CA1 region of the hippocampus, GluN2D mRNA was detected in interneurons but was absent from pyramidal neurons. By contrast, GluN2A and GluN2B were conspicuously expressed in interneurons and pyramidal neurons. In the dentate gyrus, GluN2C and GluN2D were detected in the hilar region but not the granule cell layer, in which GluN2A and GluN2B were expressed. Importantly, these higher-resolution data provided evidence that GluN2C and GluN2D become restricted