You initially had me at a loss, as you are the first person to ask me to elaborate.
Thanks for asking. It has forced me to think things through a little more and might lead to some interesting discussion.
We know that certain autosomal dominant mutations (APP/PS1) in the amyloid pathway lead to AD. Amyloid plaques in these individuals are likely a direct result of the their mutation and possibly may be the cause of disease pathology. ApoE4 is less efficient at clearing amyloid and may lead to greater deposition of amyloid. This, however, does not necessarily have to be the mechanism by which apoE4 leads to cognitive decline. Robert Mahley and colleagues are big on the toxic fragments / mitochondrial dysfunction hypothesis. Other believe it could be neurofibrillary tangles, glucose regulation, etc. Beyond that, there are clearly AD cases with ApoE3 individuals, who would have neither the autosomal mutations or apoE4. How do they get plaques and/or tangles and what leads to their cognitive decline? It's not going to be exactly the same as APP or apoE4. There may be three separate diseases that are all called Alzheimer's disease. That doesn't even go into the whole "mixed pathology" vs. "Pure AD" that is actually observed in autopsy.
I am a believer of the two-hit hypothesis for Alzheimer's disease, so I tend to think in those terms. APP/PS1 + aging are two clear hits. ApoE4 is a hit, but needs to be combined with some problem (lysosomal dysfunction, glucose dysregulation, compromised cardiovascular function, perhaps even viral infection like HSV-1). ApoE3 carriers may have two other hits. In this compromised brain, normal clearance mechanisms may not be working properly, getting to overtly similar gross pathology irrespective of the original cause. Because of this, removing plaque, as evidenced by the bapineuzumab and solenuzumab, is not a panacea. Some argue that patients were started too late, but I think that's just more of the goal post moving that commonly occurs in the amyloid field (like the oligomer hypothesis).
As I said in the previous post, the above is controversal thinking, but other than the Biogen drug data from a few months ago (which probably won't work for apoE4 carriers due to side effects), the amyloid hypothesis has little to show for itself after 20+ years and billions of dollars.