Generally, my comments are offered with the intellectual humility required when deliberating over any biomedically complex disease. More specifically, they are offered with my deepest admiration of you, one of our great pioneers who has taught and inspired us on how to pin down AD’s neurobiological complexities.

It is therefore unsurprising that I agree with nearly everything you propose in your article:

• Yes. As a recently minted drug developer, I too am in awe of how industry has persevered and how the recent crop of drugs has “hoovered out” extracellular amyloid aggregates from the brain, back to near normal.
• Yes. As a clinician, I completely agree that the clinical trials establish that extracellular amyloid aggregates are cognitively detrimental. Which is why I in good faith prescribe amyloid-clearing drugs to eligible patients.
• Yes. Although not schooled in FDA thinking, I would consider it justified to approve any drug that clears extracellular amyloid aggregates back to normal.

But no. Scientist to scientist, I simply don’t understand the justification for concluding that a convergence of findings you summarize in your paper establishes that extracellular amyloid aggregates represent a cause of AD. With respect, it seems that they have not; not yet anyway. Applying the dispassion of the scientific method, it seems more accurate to conclude that, at least for now, the findings have established that extracellular amyloid aggregates represent one defining feature of AD that has proven cognitively detrimental.

This is not merely a semantic distinction. It is at the heart of the matter, since targeting a causal mechanism promises to have the greatest therapeutic benefits. As in all neurodegenerative disease, it is AD’s neurodegenerative process, defined as the gradual loss of brain tissue, which is the prime driver of brain atrophy, and the debilitating cognitive deficits suffered by patients. It’s AD’s neurodegenerative process, therefore, that needs to be explained for causal conclusions and for therapeutics.

You are right, of course, to anchor any potentially causal conclusion on AD’s causal genes—which is now represented by the causal quartet of APP, PSEN1, PSEN2, SORL1—and relying on what defining AD pathologies they induce in model systems. It is true that they all reliably coax neurons to produce and secrete more Aβ peptides from neuronal endosomes, and that, once secreted, they end up forming extracellular aggregates. They all therefore induce AD’s defining extracellular pathology.

But you also know that they all “poison” the neuronal endosome and, subsequently, the rest of its downstream endolysosomal network, leading to another of AD’s defining pathologies—intracellular endosomal enlargement. Notably, these studies have established that this intracellular pathology can occur before, and independently of, extracellular amyloid pathology.

So, the critical question, important for causality and therapeutics (a redundancy!), can be boiled down: Which of AD’s defining pathologies induced by these causal mutations, the pathologies that occur outside or inside of neurons, triggers AD’s defining neurodegenerative process?

The argument against extracellular pathology
As shown for most of the causal genes, but also observed in sporadic AD, there is an apparent uncoupling of amyloid and neurodegeneration based on the glaring anatomical mismatch between where extracellular amyloid aggregates are first and foremost found, versus where neurodegeneration occurs first and foremost.

I have heard many counterarguments trying to explain away this perceived uncoupling between amyloid and neurodegeneration. While they have always sounded contorted and forced, they were possible; it is a complex disease, after all.

Except that the clinical trials testing amyloid-clearing drugs provide experimental evidence against aggregates causing neurodegeneration, at least for now. The trials clearly establish that aggregates are mildly detrimental to our cognition, but there are many mechanisms by which extracellular aggregates of any kind can mildly impair neuronal function without necessarily triggering neurodegeneration. And in fact, as well known, the trials have failed on readouts of neurodegeneration.

Speaking for my patients, I truly hope that amyloid-clearing drugs will ameliorate neurodegeneration when given earlier in the disease process. Because if they do, they will undoubtedly be overtly beneficial, without all the current handwringing on whether a 30 percent slowing in cognitive decline is meaningful or not.

The argument in favor of intracellular pathology
In model systems, all the causal genes induce amyloid-free endosomal enlargement, an intracellular pathology first established in “sporadic” AD brains. This amyloid-free intracellular pathology localizes to the entorhinal cortex/hippocampus, precisely where the neurogenerative process and its associated cognitive impairment occurs first and foremost.

Mechanistically, by studying causal genes, endosomal enlargement appears to be driven in part by blockage of the endosomal recycling pathway. The swelling of endosomes can easily be explained, because it is this specific pathway that is required for normal recycling of glutamate receptors and other integral proteins and lipids, and which is why the intracellular pathology has been referred to more specifically as "endosomal traffic jams."

Also easy to explain is the pathology’s ability to trigger neurodegeneration, because the recycling of glutamate receptors and other cargo are so critical for synaptic function, integrity, and ultimately for neuronal health. In fact, when modelling endosomal recycling defects found in patients, the evidence suggests that it does induce the neurodegenerative process in part by reducing glutamate receptor recycling, in the absence of extracellular aggregates.

Resolving this debate
Socrates was of course right: Only by agonistic (not antagonistic) debate do we ever get closer to a truth. Unfortunately, in biomedicine, even honest and vigorous debate can only get us so far. The ultimate way to resolve a debate on human disease is via experimental medicine, which is to say clinical trials. Future trials testing drugs that either clear out amyloid or clear up traffic jams, early in the disease process, are required to resolve whether it is AD’s extracellular or intracellular pathologies that is the prime driver of neurodegeneration.

Dennis, I will end where I began. Because of my admiration of you and your rigorous practice of the scientific method, I look forward to hearing back on this scientific analysis. I had the fortune of meeting with the great thinker and Noble laureate Danny Kahneman, before he sadly died. I recount our discussions in my recent book, where I quote how we agreed that nothing gives us more pleasure than changing our minds. Please give me this pleasure.

Sincerely,

Scott A. Small
The Boris and Rose Katz Professor of Neurology
Departments of Neurology, Psychiatry and Radiology
Director of the Alzheimer’s Disease Research Center
Columbia University