Last week’s publication in Science of a paper by Cramer et al on the effects of bexarotene in transgenic mouse models of Alzheimer’s stirred up quite the media storm, and a flurry of requests from AD families for off-label prescriptions of this retinoid-X agonist drug used in oncology. Frankly, my first reaction was one of cynicism born of…. too many mice cured, not any humans. But digging deeper, and consulting with a number of Alzheimer’s experts, my skepticism is tempered, at least temporarily. The core concept of the tactic is that RXR agonists increase APOE expression, leading to beta-amyloid clearance. The paper provided robust evidence that in these mice, both soluble and insoluble plaque levels were rapidly reduced. Very importantly, these mice also showed a rapid improvement in cognitive function, as would be predicted by the model that soluble AB interferes with synaptic function. Given that this is a drug with ample data regarding its shortterm use in humans, a compelling case exists for aggressively pursuing this as an AD candidate. However, several questions remain to be answered:
1) The predictive power of mouse models where amyloid production is hugely amplified has yet to be shown–indeed there have been numerous failures in trying to do so.
2) This does presume the primacy–or at least salience-of the amyloid model, where improving clearance of brain amyloid is seen as a key target. Again, this needs to be proven in humans.
3) Do acute changes establish longterm modification of the disease process?
4) Would bexarotene be safe when given longterm, and at what dose? (and would safe chronic dosing be sufficiently impactful on AB?)
The exciting part of the news is that some of these questions could be rapidly answered (not so much the safety of chronic dosing). The first step would be to see if bexarotene, in human AD patients, alters beta-amyloid biomarkers in what we believe to be the desirable direction. Based on the mice, one would expect to see such a shift quickly. One could first assess a sample population for biomarker changes, and then (if warranted) begin testing for cognitive changes compared to baseline–six months would be a sufficient pilot study for both the initial magnitude and durability of cognitive improvement. IP issues would complicate pharma investment in the US–but this might be a worthy use for some of the additional Alzheimer’s funding recently announced by HHS.