Expert Conversations: The Role of Amyloid in Alzheimer Disease

In this podcast, Neurology Learning Network's Alzheimer Disease Section Editor Daniel Claassen, MD, interviews David T. Jones, MD, about common ideas and misconceptions about the role of amyloid in Alzheimer disease (AD), assessing amyloid status among patients, areas of future research for the role of amyloid in neurodegeneration and AD, and more. A full transcript is provided below.

Discover more insights from your peers in our Alzheimer Disease Excellence Forum.


About the Speakers:

Daniel Claassen, MD, MS

Daniel Claassen, MD, is an associate professor of neurology, and division chief of Behavioral and Cognitive Neurology at Vanderbilt University Medical Center in Nashville, TN.

David T. Jones, MD

David T. Jones, MD, is a consultant in the Department of Neurology's Division of Behavioral Neurology, and assistant professor of neurology and radiology at Mayo Clinic in Rochester, MN.


Additional Reading:

  • Botha H, Mantyh WG, Graff-Radford J, et al. Tau-negative amnestic dementia masquerading as Alzheimer disease dementia. Neurology. 2018;90(11). doi:10.1212/WNL.0000000000005124
  • Jones JT, Graff-Radford J, Lowe VJ, et al. Tau, amyloid, and cascading network failure across the Alzheimer's disease spectrum. Cortex. 2017;97:143-159. doi:10.1016/j.cortex.2017.09.018
  • Panza F, Lozupone M, Logroscino G, Imbimbo BP. Nat Review Neurol. 2019;15:73-88. doi:10.1038/s41582-018-0116-6



Daniel Claassen, MD: Thank you for joining the podcast for the Neurology Learning Network. My name is Daniel Claassen. I'm a neurologist at Vanderbilt University. It's my delight to be with Dr David Jones, who's a consultant at the Mayo Clinic in Rochester, Minnesota.

I'll first tell you a little bit about myself. I'm the chief of the Behavioral and Cognitive Neurology Division at Vanderbilt. My research looks at translational approaches to offer therapeutic outcomes for patients that suffer from neurodegenerative diseases.

I know Dr. Jones from a long time ago when we used to be residents together. I have been excitedly watching his career go forward. He's made a lot of important and insightful advances in terms of understanding how the brain changes with neurodegeneration, how to understand networks, especially, and how networks are important.

It's an absolute delight to have him here. Thank you, Dr Jones, for joining us today.

David T. Jones, MD: Dan, thanks. It's my pleasure. It's great to be back with you, and I’ll maybe give a little bit of introduction to myself, as well.

I am a practicing clinician. I work in a behavioral neurology clinic focused on dementia. We are a tertiary referral center, and so it does tend to be a lot of young or early onset dementia. I'm also the co‑director of the Normal Pressure Hydrocephalus Clinic, and a director of Artificial Intelligence.

I have research interests in biomarker development, quantitative neuroimaging, and systems level pathophysiology of aging, cognitive aging in Alzheimer's disease and related disorders.

Dr Claassen: Excellent. You must not sleep much, but that's a start. I'll start off the conversation, and we're going to talk mostly about amyloid and what it means for clinicians, and how clinicians should think about or at least approach the ongoing discussions about clinical trial development, and even their practice.

Dr. Jones, I'll start by saying we know that amyloid's probably important in Alzheimer's disease.

When I think about the field and the way it's progressed, the 2 concepts that really come to mind is the genetic work, which especially in Down syndrome, which has really linked the role of amyloid in cognitive impairment and dementia, and secondly, the ability to monitor amyloid both through PET and NCSF markers.

Certainly, being able to see something and watch how it changes has given a potential opportunity for interventions with the hope that by modifying accumulation and plaques, that one would be able to improve therapeutic outcome.

From your perspective as a practicing clinician, can you give us a sense as to how you consider amyloid in Alzheimer's disease, both in the diagnostic workup, and then when you explain it to patients, how you talk about amyloid involved in the clinical presentation of a patient?

Dr Jones: Yeah, those are great questions, and obviously big questions in the field. I think when you hear answers from me about this, you're probably going to need some more context. I probably have a pretty nuanced view in that it's not very black and white.

I get the sense that there's people out there who think amyloid is toxic and causes Alzheimer's disease or there's people out there who think it has nothing to do with the disease. I don't know if that's a straw man argument if those people really exist, but I'm neither of those people.

I would say amyloid—is that involved in Alzheimer's disease? I'd say by definition, it is. There's no doubt or question about that in my mind. Is amyloid related to Alzheimer's disease? Absolutely. There's no doubt about that in my mind. We can at least start there, which is a firm foundation, I think.

Then, the question will be how is it involved? That's where I probably will differ. That's why I say it's probably nuanced. Do I believe amyloid is toxic? I would say no. I don't believe that it's toxic in the way that most people think about it as toxic.

I think that that's a common place for people to start having a conversation about amyloid because a lot of people are very clear that every form of amyloid is not toxic. They'll concentrate on different forms of amyloid, whether it's a monomer, protofibrils, oligos, plaques.

They'll say one of those is toxic, but the others aren't. Clearly, to actually have the amyloid hypothesis, you want one of those forms to be toxic. I'm not saying it doesn't have adverse consequences, but I don't think of it as directly toxic or any of those species in the way I think most people do.

Given that I take a systems perspective in neuroimaging, it's pretty clear to neuroimagers that you see amyloid everywhere. You see it everywhere in the brain, at least the plaques, the plaque form. It's clear that every area of the brain is not being inflicted with some toxic injury in Alzheimer's. I think that's clear.

The other thing that's clear is that it doesn't matter what form of Alzheimer's disease you have, which you know as a clinician, there's many different forms, many different phenotypes and because different areas or different systems in the brain are being affected, whether it's the visual system, or the memory system, the executive system, or the language system.

In all those patients with different clinical symptoms, the amyloid looks the same. It doesn't really vary by phenotype. The amyloid is also there before clinical symptoms arise. It doesn't correlate with clinical symptoms. These pieces of information help us. Amyloid's involved. It's not toxic. It doesn't track with cognition.

When you want to think about how it is involved, you have to also account for those pieces of information. A systems perspective is ideal at that, and it can do that.

Dr Claassen:  We'll talk about 2 things with that excellent preamble, but with the role of the systems perspective, the first thing I just want to clarify for our listeners is when you talk about systems perspective, are you referring to networks that seem to change with cognition or something else?

Dr Jones:  We'll start with the reductionist perspective and build out. Amyloid comes from APP, amyloid precursor protein. That is a synaptic protein. That protein lives at the synapse. What does the synapse do? The synapse is a core element in functional brain activity.

Embed that synapse within a collection of neurons organized into a system, which you need to have a high level of function, memory thinking, something like that. Now, the synapse, and APP sitting at the synapse, is related to large‑scale brain networks, which are required to have high‑level cognitive functioning.

When you observe changes in APP processing ‑‑ obviously, APP processing needs to be changed in order to get amyloid plaques ‑‑ you could then make a connection between what's happening at the systems level and those plaques.

That's what I mean by taking a systems perspective on thinking about what plaques are doing, why they're there, and how they're involved in the pathophysiology of a disease that attacks brain function.

Dr Claassen: Practically, if you're seeing a patient that has amnestic predominant symptoms, and they ask you, "Should I get more scans or more assessments like lumbar punctures," how do you counsel that patient in terms of interpreting either amyloid uptake on a PET scan or amyloid reductions in an LP?

How do you go about practically taking that to the patient's bedside?

Dr Jones: A lot of that depends on age. Age, because someone with memory complaints in their 50s is much different than someone with memory complaints in their 80s. We also know that becoming amyloid‑positive, the probability of that increases exponentially with age. We also know individuals with no symptoms can be amyloid‑positive.

Especially when you're thinking about older individuals, they may have amyloid positivity. That may be entirely unrelated to their cognitive symptoms. That's especially true in TDP‑43‑associated conditions. LATE disease is the new term for that ‑‑ limbic‑associated TDP‑43 encephalopathy.

That's a memory‑predominant disorder. That's where the big problem comes in here in older individuals. LATE disease protects the hippocampus for memory. Those individuals may be amyloid‑positive, and they may not be. They'll have the same clinical scenario.

We had a recent publication. We titled it ‑ we thought controversially to try to bring some attention to this ‑‑ but we called it "Tau‑negative amnestic dementia masquerading as Alzheimer disease dementia" to make the point that this is an individual who clinically looks like they have Alzheimer's disease in that they have severe amnesia, and they have dementia.

The hippocampi are very atrophic. Then you would think, based on our classical training, we'd say, "Oh, that's Alzheimer's disease." If you actually image them or look pathologically for Alzheimer's tau, you don't find it. You may find amyloid, but I would say even if you did find amyloid, it's irrelevant to their clinical situation.

Dr Claassen:  What's that age cutoff for you?

Dr Jones:  That's arbitrary. Seventy-five is a nice rule of thumb. To identify these patients, there's other pieces of information. If they're older, they have amnesia, it's amnestic dementia syndrome. You look at their brain scan, their MRI. They have very severe hippocampi atrophy. Severe.

Yet, they only have a little bit of a mild memory condition, and you don't really find any cortical signs on physical exam. If you're thinking that's a primary hippocampal problem that's only causing a very minor deficit in memory, that's probably LATE disease, especially if they had a long amnestic prodrome.

I'll usually develop that hypothesis while seeing the patients. I'll test the hypothesis by looking at the scan and seeing if it looks like I think it should in that scenario, the hippocampus.

I'll also start asking the patient about long‑term semantic information, which is usually lost in these people with LATE disease, salient events that they should remember, like details about 9/11, who Osama bin Laden is and whether he was involved.

Usually, the family then sees that if you can't remember those types of details, something's wrong, whereas when they came in the office, they would say, "He just can't make new memories." They say the old ones are fine, but you question specific old memories, and they're gone.

This starts to build up a picture where I just don't care about amyloid anymore. I wouldn't get amyloid imaging. I wouldn't get a CSF amyloid study because even if it was positive, it wouldn't change my mind.

Dr Claassen: That's really great. Maybe a younger patient that has amnestic symptoms, and you look at the scan, and you see maybe some parietal atrophy, the amyloid might be helpful, but it may not change your mind also based on what you're seeing on the scan.

Dr Jones: Given how rare amyloid positivity is in younger individuals, that means something to me, especially if they're under the age of 60. If you're under the age of 60, you're amyloid‑positive and you have clinical symptoms, it's likely Alzheimer's disease. You may have incidental amyloid positivity in people under the age of 60, but they don't have symptoms usually.

You'd have to have a rare confluence of events to have an incidental amyloid case under the age of 60 present as clinical symptoms from frontotemporal dementia or some other neurodegenerative condition. That's pretty rare. Your pretest probability changes a lot based on those 2 facts, age and clinical symptoms, that are present.

Dr Claassen: That's informative. There's been so much news about amyloid‑targeted therapies for Alzheimer's disease. A lot of people have written about it and offer their opinions.

If you just think about the potential for modifying amyloid and changing the inexorable degeneration that may or may not happen, how do you think about amyloid in terms of clinical trials, and how neurologists should think about interpreting these trials?

Dr Jones: I have to think about it in terms of a pathophysiologic model. I have to otherwise, I don't know how else to think about it or at this point. That's why I've been working hard on developing the cascading network failure model to figure out how amyloid fits in.

Dr Claassen: Can you tell us a little bit more about that that model, the cascading network model? I think that's really interesting.

Dr Jones: Going back to the synapse, you can think about APP as a synaptic protein. Instead of thinking about this is amyloid disease, you can think about this as an APP‑processing disease, or you can think about it as a synaptic disease because APP is a key synaptic protein.

You can think about this as Alzheimer's, as a synaptic dyshomeostasis. If you can embed that synapse in a large‑scale system of many synapses, of many neurons, dyshomeostasis in that large‑scale system will lead to dyshomeostasis at the synapse, which will lead to dyshomeostasis of APP processing, which will lead to an amyloid plaque.

In the same way, your thoughts can exist at this level. The way your brain coordinates those dynamics in the large‑scale system to maintain homeostasis can influence processing of APP at the synapse.

Considering what's happening at the system level, which we know there's decreases in connectivity in the back of the brain, there's increases in connectivity in the front of the brain that happens with aging. We see those same changes in Alzheimer's disease. We've associated those with amyloids and tau.

If you incorporate what's happening to large‑scale systems to maintain normal functioning and see what the consequences may be for synapses and vice versa because it's obviously bidirectional ‑‑ you don't have systems without synapses, there's no point in synapses if you don't have systems ‑‑ it becomes complex.

That's not just throwing your hands up. That's a technical term, complex systems, where you relate these different scales and complex dynamics. To make it simple, there's some age‑related changes to maintaining normal homeostasis in brain networks. That's probably a tau‑associated process in Alzheimer's disease.

Then the adaptive responses to maintain homeostasis change synaptic processing and hubs in the brain. That facilitates the formation of amyloid plaques. Amyloid plaques or APP processing can lead to changes in the synapses that can be a buffer.

Once you kind of run out of this adaptive capacity, the original tau‑associated process is released. It's allowed to move forward.

That's the nutshell of the cascading network failure hypothesis. You have an original focal problem, maybe within a module that's more tau‑associated. You have distributed global adaptive responses that takes place over a decade. Once the adaptive capacity is exhausted, the original degeneration, which was being adapted for, is free to move forward.

That also explains the global patterns. It explains the spatiotemporal differences between amyloid and tau. It explains why you can be amyloid positive and have no symptoms. It also explains why you never see tau take off until you get amyloid ceiling. It explains a lot of things which are unexplainable with other models.

Dr Claassen: That's a really interesting idea. One of the areas that a lot of people have spent time looking at, and certainly our group has looked at, is this idea of amyloid clearance and potentially the role of a glymphatic network and things like that.

Do you see a role for understanding CSF flow clearance of proteins and such in that model, out of curiosity?

Dr Jones: Anything that could affect homeostasis of APP, it could play a role. I would say it's likely a contributor, and I'd say you absolutely can have Alzheimer's disease with normal lymphatics. I have no doubt in my mind about that, but could abnormal lymphatics interact? Yeah.

Dr Claassen: If amyloid was available to any neurologist as part of their workup, how would you counsel them to order it? Would you suggest first have under a psychology exam clinical assessment and then order it? There could be cases where you'd want to get it before you see the patient.

I get a lot of questions from the community about ordering this, and I just wondered if you could provide some advice to folks on that issue.

Dr Jones: Right now, this is in the pre‑amyloid treatment age it's very different than in a post‑amyloid treatment age, I would say. Right now, I would not recommend anyone without specialty expertise ordering a blood‑based amyloid biomarker outside of the specialist context. I would discourage that.

Dr Claassen: You feel like with all the emphasis on these blood‑based biomarkers, there's going to be some advertising.

Dr Jones: Exactly. I think there need to be a lot of education about them, what they mean, what they don't mean. Like any test, you got to be Bayesian about it. You got to have some pretest probability to interpret it. I'm worried on the interpretation of some of these things without the proper clinical context.

Dr Claassen: This has been really helpful. One of the other questions I wanted to ask you is what do you think the scientific community needs to answer regarding the cascade model? I guess I should ask that question a little differently. What methods do you think we need to develop to get answers to that cascade model that you suggested?

Are there better longitudinal studies that we need to do? Are there method development, imaging studies that we need to pursue? How do you see the future in that regard?

Dr Jones:  Fundamentally, it's a model about neurodynamics. We need better biomarkers of neurodynamics. What I mean by neurodynamics is the large‑scale neurodynamics, the functioning of the brain over time.

To put it very easily, maybe easily in my mind, but there's difference between the mind and the brain. When you die, the brain stops functioning and then the mind goes away. What went away was neurodynamics.

Neurodynamics is what I'm interested in. We need biomarkers of those across aging. We need to know what healthy neurodynamics are. We need to know what unhealthy or risky ones are. Then, we need to relate those to molecular pathophysiologic measures.

That'll help us understand, test hypotheses made by the cascading network failure model, but also will guide mechanistically‑driven therapies and outcomes. Better biomarkers for neurodynamics is what we need.

Dr Claassen: You think those will be imaging?

Dr Jones: Another question would be to have good biomarkers of neurodynamics, you actually have to have a good model of neurodynamics.

Dr Jones: That's another prerequisite to moving forward is knowing what you're looking for, and then you'll know how to look for it.

I think electrophysiology is probably the better place to go. Right now, we have more well‑developed theories and understanding of what's happening from a functional MRI perspective. That's going to be a useful tool to transition to electrophysiology.

In the clinical phase, FDG is key, the key functional measure in the clinic, of the consequences of abnormal neurodynamics manifesting in metabolism. FDG is a key modality for clinical evaluation of cognition.

Dr Claassen: We've seen that. I think others have tried to use arterial spin labeling and things like that to try and reduce some of the radiation exposure. Certainly, looking at FDG can be very helpful with understanding metabolic changes. I don't know if you used ASL? Do you have any ideas on that for another option for people? We get that question as well.

Dr Jones: I use FDG routinely in the clinic. I see its power, does not replicate what I see with ASL in the clinic that I've seen so far. Some big issues with ASL, it requires very sensitive to the acquisition and the processing and artifacts and also vascular change in the brain.

If you’ve got a stenosis or anything like that, that's going to influence your measure, whereas FDG has none of those problems.

Dr Claassen: This is really informative. None of us are going to ask more questions out of this. At the end of this podcast, we'll put the links to some papers that folks can read about these models and some of these questions.

David, thank you very much for joining us today. This is really useful. I look forward to catching up again to see if we can answer some more questions later on.

Dr Jones: We'll probably end up asking more.

Dr Jones: Or being more confused. That's typically what happens after a conversation with me.

Dr Claassen: No, this is good. This is good. Thank you all for listening. Thank you for being part of the Neurology Learning Network, and we will look forward till next time.