A recently published paper reported that combination of three diabetes drugs (GLP-1, GIP, and glucagon) improved memory in a mouse model of Alzheimer’s. This is quite remarkable, but not shocking.

There are several competing theories of Alzheimer’s disease. There is the all-but-discredited theory of beta amyloid. I would like to still believe in this theory, having been a proponent of this theory for several years at Elan. It is a compelling theory for several reasons. First, the familial forms of Alzheimer’s disease have mutations that increase beta amyloid. Second, you can cause Alzheimer’s by injecting beta amyloid into the brain. Third, transgenic mouse models targeting beta amyloid can reproduce some of the symptoms of the disease. But unfortunately, data from multiple studies have shown that the theory is incorrect. It is difficult to admit that so much beautiful science (much of it done at Elan) can be so wrong but I think it may be.

Unfortunately, the beta amyloid theory is deeply ingrained. A very bright researcher I know submitted a grant to the NIH with a very compelling idea about a non-beta amyloid cause of Alzheimer’s disease. It was rejected because the reviewers “didn’t see how the proposal fit into the beta amyloid theory.” Science in case may proceed one funeral at a time, as Max Planck once noted–in some fields, only when the old scientists retire or die can the better and fresher ideas take the stage.

Then there is the infection theory. This is another compelling theory. Beta amyloid appears to have antimicrobial effects, so it may be a immune defense protein. And, in amyloid-mediated diseases like mad cow disease and CJD, it appears that a pathogen rather than prions may be the causal agent. There have been some elegant work done by the Manuelidis lab at Yale on this.

And there is the diabetes theory. People with diabetes have double the risk of Alzheimer’s than non-diabetic people. There are some people who believe that Alzheimer’s is just another manifestation of metabolic syndrome, and they are calling Alzheimer’s disease Type 3 diabetes. In other words, it’s just a late manifestation of metabolic syndrome.

I think they may be onto something. I’ve always thought it odd that insulin degrading enzyme is also the enzyme that degrades beta amyloid. I’ve also thought it odd that Type 2 diabetes was associated with amyloid deposits in the pancreas. And I’ve thought is even odder that people ignore the amyloid in diabetes but focus on it to the exclusion of almost everything else in Alzheimer’s disease.

But even odder, is the fact that if you give intranasal insulin to patients with cognitive impairment, their cognition improves. Seriously, before you take your next standardized test, try a shot of insulin nasal inhaler. It’s one of the most effective nootropics, apart from nicotine. But oddest of all? This improvement in cognition is related to what kind of ApoE allele you have. Yup, the same ApoE that has an enormous impact on cardiovascular risk and the risk of Alzheimer’s. ApoE4 carriers have an enormously increased risk of Alzheimer’s disease. It also raises the risk of cardiovascular disease. The patients with ApoE4 allele, insulin had the least effect.

So is it surprising that diabetes drugs help cognition? Yes, a little surprising but not shocking. We already know that insulin helps cognition. And the evidence for a protective role GLP-1 in Alzheimer’s disease has been accumulating.. For example, mice that are treated with streptozocine develop diabetes, as is well known. But they also develop Alzheimer’s beta amyloid plaques. Remarkably, this can be prevented with GLP-1. But it does suggest that the Type 3 diabetes theory of Alzheimer’s may be worth a close look.