The amyloid hypothesis has dominated Alzheimer’s disease research for 25 years and generated major advances for the field. But as noted by Bart De Strooper and Eric Karran in the February 11 Cell, some of that new knowledge does not sit well with the hypothesis. How to explain the decades-long prodromal period of AD when amyloid plaques accumulate but neurons survive and cognition remains intact? Where are the links between proposed toxic species of Aβ and neurotoxicity? De Strooper and Karran emphasize the need to reconcile the amyloid hypothesis with the complex cellular makeup of the brain. They posit that after toxic species of Aβ and tau begin to accumulate, a multiyear cellular phase ensues during which aberrant neurons, glia, and vascular cells permanently alter the brain. Better systems biology approaches, in particular single-cell resolution of the perturbations during this phase of the disease process, could prove invaluable in stopping the irreversible, progressive neurodegeneration that occurs in AD, de Strooper and Karran claim. 

De Strooper and Karan debated their ideas with Todd Golde, David Holtzman, and Beth Mormino on March 24th.

We thank Cell Press for graciously giving Alzforum readers free access to the review.

Please contact us for technical questions prior to the event.




By Tom Fagan

The amyloid hypothesis has survived major advances in Alzheimer’s disease research over the last 25 years but, given some of those advances, this central concept of AD pathogenesis may have to evolve. Notably, no agreed-upon mechanistic link has emerged between Aβ and neuronal toxicity, and the decades-long incubation period during which plaques accumulate but neurons continue to function remains hard to reconcile with the neurocentric basis of the hypothesis. In their review in the February 11 Cell, Bart De Strooper and Eric Karran argue for a more holistic approach. They expand the original, linear view of the amyloid hypothesis, and instead conceptualize Alzheimer’s disease in terms of three sequential phases (see diagram below). In the initial biochemical phase, which may last about a decade, Aβ accumulates as per the amyloid hypothesis, as do hyperphosphorylated tau, plaques, and tangles. Next comes a decades-long cellular phase, during which neurons, glia, microglia, and vascular cells engage in feedback loops of compensatory activity, which slowly chips away at synapses and functional circuits. This, eventually, leads to the clinical phase expressed as the symptoms of dementia.

The Three Phases of AD. Complex interplay among genes, molecules, cells, and circuits contribute to the three phases of Alzheimer’s [Image Courtesy Cell, De Strooper.]

De Strooper and Karan hypothesize that the cellular phase starts with reversible changes as cells respond to proteotoxic stress. As the brain compensates, permanent changes slowly accrue. While much is known about Aβ generation and aggregation, this second phase of cellular defense and warfare remains much murkier to science. The authors claim that a better understanding, particularly of the cellular phase of AD, could yield valuable insight into the disease process, and provide a stronger basis for the development of targeted therapeutics. For example, single-cell approaches could help map changes in gene expression in multiple cells in parallel. This could uncover spatial or temporal differences in how cells respond, even among subgroups of the same cell type. The field at large would gain an invaluable resource if this were done by Braak stage in different cell types throughout the brain and made publicly available. This and other projects exploring the cellular phase of AD could propel the field’s research into the 21st century, the authors claim.



  1. De Strooper and Karran’s thoughtful review takes us past the initiating events in AD, to think about the entirety of the pathogenesis. With all the new genetic risk factors we have discovered, this is an entirely appropriate development. We have been thinking of AD as a neuronal disease, when these new genetic risk loci, and indeed an examination of the pathology, shows it is a tissue disease. This review is the first attempt to fully incorporate these new insights into the amyloidogenic framework.

  2. Left unexplored in this otherwise inclusive analysis of the pathogenesis of AD is any discussion of the factors that initiate the various pathogenic processes that precede and lead to the accumulations of “toxic” amyloid and tau. A case can be made for oxidative modifications of DNA and RNA, DNA strand breakage and aberrant repair, and the still-unaccounted consequences of telomere senescence and aging. Early white-matter changes, which seem to precede the toxic accumulations of amyloid and tau, can be explained by vascular damage of large and small blood vessels that result in both ischemic injury and glymphatic drainage problems. Identifying and dealing with these primary causes is a path yet to be vigorously explored.

  3. The cellular phase, defined by De Strooper and Karran, appears to largely represent intracellular and intercellular interactions responsible for AD development. They probably involve both gain of toxic functions and loss of normal functions. In my understanding, De Strooper and Karran intend to incorporate into their four-dimensional scheme all the major cellular reactions that lead to neurodegeneration. This is why we need to understand cellular reactions in vivo, possibly on a single-cell basis. I guess that complexity of the brain will make us work for a long time. Also, some reactions probably constitute “AND” circuits, others “OR” circuits. In this respect, recent mathematical neuroscience, the advancement of which is incredible, may help untangle the complexity issue.

    I am not, however, a big fan of systems biology because it depends on past publications, many of which are unreproducible or irrelevant. The best way probably would be to observe humans from their preclinical stage for 25 years in biochemical, cellular, and clinical terms, and this is impossible. The second-best classic way is to reconstitute all the major mechanisms in animal models. We have to keep in mind that all the rodent models that accumulate Aβ are not perfect models of AD; they are, at best, just models of preclinical AD, which might develop symptomatic AD if they lived for 25 years, by when many of us will have retired. The answer still remains behind the veil.

Make a Comment

To make a comment you must login or register.


Other Citations

  1. contact us

External Citations

  1. review

Further Reading

No Available Further Reading

Primary Papers

  1. . The Cellular Phase of Alzheimer's Disease. Cell. 2016 Feb 11;164(4):603-15. PubMed.