Background

Rosiglitazone is an antidiabetic medication that lowers blood glucose by sensitizing pancreatic target cells to insulin. Rosiglitazone belongs to the thiazolidinedione class of peroxisome-proliferator activated receptor γ agonists; PPARγ is a nuclear receptor predominantly expressed in adipose tissue, whose activation influences expression of genes involved in glucose and lipid metabolism. The FDA approved rosiglitazone for treatment of Type 2 diabetes in 1999, but side effects soon started being reported (Sep 2003 news) and their use was restricted in 2010 after having been linked to heart attacks and strokes (see New York Times article).

PPARγ has come to be a target of interest for Alzheimer's drug development because of an overlap of metabolic disease and Alzheimer's disease risk factors. Cell-based and animal studies have implicated PPARγ, in particular, as playing a role in neuroinflammatory processes in neurodegenerative conditions. For example, PPARγ activation has been shown to modulate the microglial response to amyloid deposition in such a way that it increases Aβ phagocytosis and decreases cytokine release (see Dec 2012 news story; Mandrekar-Colucci et al., 2012). 

Rosiglitazone has been variously reported to improve behavioral outcomes, normalize brain glucocorticoid receptor levels, activate wnt signaling, and decrease insoluble Aβ42 levels in mutant human APP/PS transgenic mice (Pedersen et al, 2006; Escribano et al., 2009; Toledo and Inestrosa, 2010; O'Reilly and Lynch, 2011). It was also reported to stimulate generation of mitochondria in mouse brain, and to protect synapses in hippocampal slice cultures against Aβ oligomer toxicity by way of increasing the number of mitochondria in neuronal spines and dendrites (Strum et al., 2007; Xu et al., 2013). Anti-inflammatory and anti-tau hyperphosphorylation effects have also been proposed (Xu et al. 2014; Yu et al., 2014).

Findings

A dozen clinical studies have evaluated rosiglitazone in AD.

Starting in 1999, a pilot study at the Veterans Affairs/University of Washington tested a six-month course of 4 mg per day of rosiglitazone or placebo in 30 people with amnestic MCI or mild AD. It reported a treatment benefit on recall and attention, as well as stable plasma Aβ in the treatment group versus declining plasma Aβ in the placebo group, claiming a slowing of disease progression (Nov 2002 news; Watson et al., 2005).

Subsequently, GSK ran four Phase 1 trials in a total of 142 patients to assess drug interactions, an extended-release formulation, and pharmacokinetics. From May 2004 to July 2008, GSK conducted a multicenter Phase 2 study in North America and the United Kingdom. It evaluated the effect of a one-year course of 8 mg per day of rosiglitazone on global and regional functional brain activity and cognition in 80 people with mild to moderate Alzheimer's disease. This trial reported an early increase in glucose metabolism, but no evidence for a slowing of progression at one year (Tzimopoulou et al., 2010).  

From January 2004 to May 2005, GSK ran a large Phase 2 trial at 67 centers in Europe and New Zealand, comparing 2 and 8 mg of rosiglitazone per day to placebo in 511 patients with mild to moderate AD for their ability to change baseline ADAS-cog or Clinician's Interview-Based Impression of Change plus caregiver input (CIBIC+) levels. The trial missed its primary endpoint, but a subgroup analysis found a treatment benefit on the higher dose in ApoE4 non-carriers, prompting a pharmacogenomic Phase 3 program (Sep 2006 news; Risner et al., 2006).

In July 2006, GSK started two 54-week Phase 3 studies of add-on therapy in people with mild to moderate Alzheimer's disease. One compared 2 and 8 mg per day of rosiglitazone plus donepezil to donepezil alone in 1,496 participants, the other compared the same doses of rosiglitazone added to any of the three cholinesterase inhibitors approved for AD in 1,485 participants. Both these multinational trials stratified participants into ApoE4 carriers and non-carriers. ADAS-cog and CDR-SB served as co-primary outcome measures. Both trials were negative. Rosiglitazone showed no effect on cognition or function with either dose, or in any ApoE genotype (Harrington et al., 2011).

Between February 2007 and September 2008, a third Phase 3 trial compared a six-month course of either 2 mg/day or 8 mg/day rosiglitazone monotherapy to donepezil treatment in 693 people with mild to moderate AD. The ADAS-cog and CIBIC+ input were co-primary endpoints. This trial, too, was negative, with no significant difference between the groups on either outcome (Gold et al., 2011). In all Phase 3 trials, peripheral edema was the most significant adverse effect.

GSK subsequently terminated two open-label extension trials, and discontinued evaluating rosiglitazone for an indication in Alzheimer's disease.