Amyloid-β is an anti-microbial peptide, a part of the innate immune system's attempt to disrupt the activities of infectious pathogens. Some data suggests that Alzheimer's disease, characterized in its early and preclinical stages by slow aggregation of misfolded amyloid-β in ever larger amounts, is driven by persistent infection. It is by no means certain that this is the case, but it does place the aggregation of amyloid-β in a somewhat different light than was originally the case, when it was thought of as molecular waste and little more.
Given that amyloid-β is performing a useful function, reducing or eliminating its production is probably a bad idea - and indeed this idea was attempted and made patient outcomes worse. The right way forward in the matter of amyloid-β is most likely periodic clearance of the harmful aggregates or harmful excess elsewhere in cells, a goal presently complicated by the failure of clearance via immunotherapies to produce patient benefits in the clinic. This may be because the wrong forms or locations of amyloid-β were targeted, or because amyloid-β ceases to be the primary pathology in later stages of the condition, when chronic inflammation and tau aggregation drive one another in a feedback loop that kills neurons and leads to death.
Neurons produce Abeta as an anti-microbial peptide (AMP), a way to protect themselves from microbial assailants. When they come in contact with a pathogen, molecules of Abeta bind to the intruder, which triggers them to stick together into aggregates. Trapping the brain bugs in a sticky web allows Abeta to deactivate the microbial raiders, protecting the brain from infectious assault. With this model, a number of things that scientists have been reporting for years suddenly start to make sense. For one thing, it's long been known that the complement system is activated in the early stages of Alzheimer's disease. The complement system is a part of the innate immune system that directly destroys pathogens by tearing open their membranes, and it was already known to be activated by other AMPs. The model also explains why proteins that are part of the complement system are often found bound up with Abeta plaques in the brain.
If Abeta is an AMP, it also reframes the role of inflammation in the aging and Alzheimer's brain, and the associated activation of brain-resident immune cells called microglia. Microglia are like the macrophages of the brain, gobbling up particulate matter, cellular debris, and other harmful materials in the brain - including, importantly, Abeta - and digesting it in their lysosomes. Microglia have receptors on their surfaces that cause them to spring into action when they get a whiff of activated complement proteins, and Abeta causes dormant complement protein precursors to be converted into their active forms. In the Abeta-as-AMP model, this becomes an elegant host defense system: Abeta is released, traps a marauding microbe in a self-aggregating web of proteins, and then activates complement to help finish off the enemy and to recruit microglia to clean up the battlefield.
This sequence protects the brain from these toxic materials in the short term - first from the infectious intruder, and then from Abeta itself. Abeta is produced in the short term as an emergency response to microbial marauders; microglia are then activated and recruited to clear the dead pathogens and aggregated proteins out of the brain so that they don't cause harm of a different sort. So long as this cycle is executed flawlessly, the brain remains protected from threats and sustains function. But none of these processes are perfect, they leave behind a few microbes here ... a few protein aggregates there ... and a few dysregulated microglia in another corner. Meanwhile, other aging processes make it increasingly difficult to close the loop on the cycle of releasing and aggregating Abeta, destroying pathogens, and recruiting microglia to clean up the battlefield afterward.
Abeta defends the brain against microbial invaders by forming aggregates that capture and neutralize them. Once they've already carried out the attack, the whole snarled-up mess - Abeta polymers, dead microbes, and complement proteins - serves no further purpose and can be toxic to the brain. So Abeta that is cleared out after becoming aggregated has already finished serving a useful purpose, and is mere battlefield rubble that must be safely swept away to help rebuild the neighborhood.