Cognitive Labs

Researchers at Takeda and Zinfandel Pharmaceuticals have found that the brain functional connectivity of rats shows better improvement with low doses of pioglitazone than with higher doses of the drug. The data, presented at the Alzheimer's Association International Conference this week, suggest that the drug – currently sold by Takeda as a treatment for type 2 diabetes under the brand name Actos – may be given to Alzheimer's patients in clinical trials at much lower doses than are administered to diabetics, thereby reducing their risk of experiencing significant adverse reactions.

The dose finding study brings the team a step closer toward a Phase III trial to test pioglitazone's ability to delay the onset of mild cognitive impairment in a genetically selected subset of patients deemed to be at high risk for developing Alzheimer's disease within five years.

In the study, called BOLD, researchers led by Karen Asin of Takeda looked at the pharmacokinetics and pharmacodynamics of pioglitazone in rats to assess if it increased glucose and oxygen utilization in certain regions of the brain. "Using a method for looking at dose determination for getting a brain effect, the surprising result is that the lower doses were better than the higher doses of [pioglitazone]," Allen Roses, professor of neurology at Duke University's Department of Medicine, Zinfandel CEO, and an author in the BOLD trial, told PGx Reporter.

In the BOLD study, Roses and colleagues gave rats doses of pioglitazone that would be equivalent to human doses of 1.5 mg to 12 mg. "Those doses are well below the doses that were used for the 22 million man years of treatment of diabetes," said Roses. The Actos label recommends starting diabetes patients at 15 mg/day and increasing the dosage as needed up to 45 mg/day.

Rats in the BOLD trial were treated at baseline with citric acid and then either continued to receive citric acid or one of four doses of pioglitazone (0.04, 0.08, 0.16, or 0.32 mg/kg/day) for seven days. One group of control and pioglitazone-treated rats received MRI scans after the second day of dosing, while all rats in the study received MRI scans after the seventh day. The researchers analyzed 57 regions of interest in the rats' brains.

Roses and his team found that 17 brain functional connections were significantly affected in the rats after they received pioglitazone for seven days. The affected regions of the brain showed improved functional connectivity after two days of receiving the drug. Compared to the rats in the control arm, two functional connections demonstrated "significant changes" in animals treated with 0.8 mg/kg/day of pioglitazone, which translates to a 3 mg/day dose in humans.

Based on these results, Asin et al. concluded in a poster presented at AAIC that pioglitazone doses that are lower than those used to treat type 2 diabetes can lead to positive changes in brain functioning in non-diabetic rats.

"Although the exact molecular mechanisms through which pioglitazone provides neuroprotection are unknown, pioglitazone stimulates mitochondrial biogenesis, improves metabolic function, reduces inflammation, and elicits beneficial changes in Aβ homeostasis through agonism of PPARγ receptors located in the brain and periphery," the investigators wrote in the abstract. Results from the BOLD study "support the suggestion that thiazolidinediones may be a useful approach for addressing cognitive deficits associated with neurodegenerative diseases, including Alzheimer's disease."

The BOLD study is part of Takeda's development program for low-dose pioglitazone as a treatment to delay the onset of mild cognitive impairment from Alzheimer's in a pharmacogenetically defined patient subset. Last year, Takeda began collaborating with Zinfandel, a company started by Roses, to conduct a study in which normal subjects between ages 68 and 83 will be randomized to receive either pioglitazone or placebo.

Previously conducted animal and clinical studies have shown that thiazolidinedione drugs, such as pioglitazone, improve cognition and dynamic functions of neuronal mitochondria in Alzheimer's disease in murine and human-derived neuronal cell culture models. Roses' team has studied pioglitazone's safety record for more than ten years while the drug has been used at higher doses in humans. Using their understanding of the drug's mechanism of action, the researchers have designed a Phase III trial in which participants will be given either placebo or a new formulation of low-dose pioglitazone only if they are deemed to be at high risk for developing Alzheimer's within five years.

Disease risk will be assessed with the help of a companion diagnostic that factors in the participant's age, APOE4 status, and TOMM40 status in order to predict the likelihood of developing mild cognitive impairment due to Alzheimer's. Those who fall in the low-risk category will be randomized to one of two placebo arms (PGx Reporter 01/12/2011).

Roses has published research showing that varying lengths of the TOMM40 rs10524523 poly-T polymorphism — located at intron 6 of the TOMM40 gene and linked to APOE ε3 and APOE ε4 polymorphisms — can be used to craft a three-allele risk prediction system for gauging the age of onset for Alzheimer's. Independently but simultaneously with the Phase III pioglitazone trial, researchers will also validate the genetic prognostic test for gauging which patients are at high risk for Alzheimer's-related mild cognitive impairment.

The rationale for using drugs for glucose control, such as pioglitazone, to treat people at heightened risk for getting Alzheimer's comes from early PET studies of the parts of the human brain associated with the disease, which showed that decreased glucose utilization may be a function of the APOE 4/4 genotype. In one study, published in 1996, Eric Reiman and research collaborators at Arizona's Aging and Disability Resource Center compared 11 normal people with the APOE 4/4 genotype against 22 people who were normal and had the 2/3 or the 3/3 genotype. They noted in that study a consistent decreased glucose utilization in multiple areas of the brain in normal people (of average age 50 years) with the APOE 4/4 genotype compared to those with no APOE 4 allele. Following this, researchers have consistently replicated decreased glucose utilization using PET studies in APOE4 carriers.

This decreased glucose utilization in genotypes associated with Alzheimer's "led us to start looking at [the diabetes treatment] rosiglitazone," Roses said. At the time, however, based on measurements of cerebral spinal fluid, researchers thought that thiazolidinedione diabetes treatments such as pioglitazone and rosiglitazone, sold by GlaxoSmithKline under the brand name Avandia, didn't get into the brain.

"What we now know is that [these drugs] get into the brain very rapidly and it gets cleared very rapidly, and we know this from the BOLD study,” Roses said. "The understanding that pioglitazone increases oxygen utilization in specific areas within the brain made it possible to use this technique in order to examine how that effect is accomplished, meaning do you have to go higher with the dose or do you have to go lower?"

With guidance from the BOLD data that a lower dose of pioglitazone had the best impact on brain functional connectivity, Takeda and Zinfandel are exploring different dosing and formulation strategies in humans. Although the Phase III study of pioglitazone in Alzheimer's patients was originally slated to start in October, that trial will begin after the researchers settle on the dose using human BOLD imaging studies.

The investigators are testing several low doses of the drug in humans to figure out the lowest dose that can be given to clinical trial participants over the five-year study period to minimize the risk of adverse reactions. Heart failure and bladder cancer are some of the serious ill effects that have been reported in diabetes patients receiving pioglitazone at higher doses and over prolonged periods.

Roses expects the Phase III study to begin by the end of this year or early next year.

Learning from Avandia's Failure

For some time now, researchers have theorized that Alzheimer's disease is a type of "diabetes of the brain." People whose brains have impaired ability to metabolize cerebral glucose often exhibit cognitive impairment and researchers feel this may be a warning sign of Alzheimer's.

Roses began following the effect of thiazolidinedione drugs on Alzheimer's disease while he was senior VP of genetics research and pharmacogenetics at GSK from 1997 to 2008. The company about a decade ago was testing out the "brain diabetes" Alzheimer's theory with Avandia, which was its best-selling diabetes drug before a study in 2007 linked the treatment with an increased risk of heart attacks.

Based on a Phase IIB pharmacogenetic-assisted trial that concluded in 2005, GSK researchers found that APOE4-negative Alzheimer's patients garnered a cognitive benefit from Avandia, but those studies did not pan out in the follow-up Phase III trial. In 2009, GSK reported that in a six-month study involving 553 patients with mild to moderate Alzheimer’s disease, normal diabetic doses of Avandia failed to improve cognition or behavior in people who were APOE4 non-carriers.

Since his departure from GSK, Roses' work exploring the relationship between age of Alzheimer's onset, TOMM40 lengths, and APOE4 status has yielded some clues as to what may have gone wrong in those earlier attempts.

The first Phase IIb trial at GSK to test rosiglitazone's effect in Alzheimer's disease patients by APOE carrier status didn't find a difference in the effect of the drug in APOE4 carriers. However, in non-carriers of APOE4, GSK researchers saw increased glucose utilization with 2 mg, 4 mg, and 8 mg doses of rosiglitazone. "This was very surprising," Roses recalled. "At the time, we thought the brain was going to be harder to get into and [in the literature] the drug was thought to be impermeable to the blood-brain barrier."

Then, again, in a Phase III study involving rosiglitazone in Alzheimer's patients there were signals that lower doses of the drug might have a greater impact than higher doses. When patients were given either 2 mg or 8 mg of the drug, the p-value for those who received the 2 mg dose was 0.07, very close to reaching statistical significance, while the p-value for the 8 mg dose wasn't even close to being significant.

However, it is worth noting that in those trials, a large portion of the Phase III study cohort was from Asian countries, predominantly China, a population that carries APOE4 at much lower frequencies than Caucasians. The Phase IIb study, comparatively, was done in all Caucasians.

"So, if you were recruiting patients with mild to moderate Alzheimer's disease correctly, you'd expect to have a lower frequency of APOE4s in the Chinese population. In fact, in those [Phase III] study populations there was an increase in the proportion of APOE4," Roses said.

It has been suggested that some investigators in the trial were recruiting Alzheimer's patients by APOE4 carrier status, resulting in a much higher proportion of APOE4-positive patients than is seen in the general population. If that was true, then researchers "averaged out the possibility of seeing an effect [of rosiglitazone] in the APOE4-negatives," Roses noted. "Still, the hint was there that the low dose of rosiglitazone might have had an effect." This finding was published three years after Roses left GSK in 2010 in Dementia and Geriatric Cognitive Disorders.

Only after the subsequent discovery of TOMM40 in 2009 and its genetic relationship to Alzheimer’s disease age of onset distributions could these data be re-modeled. Roses requested the DNA samples for the trial from GSK two years ago, but the company did not make the samples available for retesting for TOMM40 polymorphisms.

Seeking the Lowest Dose

Still, going off the early signals from the Avandia studies, Roses and his team crafted the rationale for the BOLD rat trial. The data suggest that pioglitazone at doses much lower than those tested in the rosiglitazone trials may improve brain functional connectivity in people at heightened risk of getting mild cognitive impairment from Alzheimer's.

Takeda and Zinfandel have not yet announced which pioglitazone doses they will use in the Phase III trial, but the companies are in the process of conducting another study, called the Confirmatory BOLD trial, to test even lower doses of the drug in humans.

In the rosiglitazone study published in Dementia and Geriatric Cognitive Disorders, researchers found that Alzheimer's patients experienced fewer toxicities with the 2 mg dose. So, when it comes to looking at pioglitazone's impact on mild cognitive impairment due to Alzheimer's, Roses and his team are interested in seeing how low they can go in dosing the drug.

"When you do a drug study, you're usually trying to insure efficacy so you can get it through the FDA, so you push the drug to almost toxicity in order to get the clinical effect. But in terms of what's the lowest dose you can get for the drug, once it's on the market, no one's going to be studying the drug in that setting," Roses said. "But if you go lower, you usually can get rid of a lot of side effects."

As diabetes treatments, the labels for both pioglitazone and rosiglitazone include a class warning noting that thiazolidinediones "cause or exacerbate" congestive heart failure. Due to the US Food and Drug Administration's growing concerns over this adverse drug reaction, the agency restricted the sale of Avandia last year to only those patients who don’t respond well to Actos.

Meanwhile, based on an interim analysis of a study that found that patients taking Actos for longer than one year may have an increased risk of bladder cancer, the FDA last year updated the drug's label to warn patients that there "may be" a treatment-related increased risk for the disease. France and Germany have asked doctors to stop prescribing the drug to diabetes patients.

Patients at heightened risk for Alzheimer's disease may be willing to accept a different risk/benefit profile from these drugs than diabetes patients since there is a critical need for a safe drug to treat or delay Alzheimer’s. Still, if the risk of adverse reactions are confirmed after completed analysis, it could make it difficult for Takeda to promote pioglitazone as a delay-of-onset treatment for Alzheimer's.

Recognizing this, Roses and his team are trying to find the lowest possible dose that will enable Takeda to study the drug in the Phase III trial in patients with an acceptable safety profile without compromising its efficacy. There was additional experience at GSK with a 2 mg dose of rosigtitazone in other indications, and in those studies the rate of treatement-related side-effects were markedly diminished.

"In some ways, the situation [with thiazolidinediones] is reminiscent of what happened with birth control pills," Roses said. When those drugs first came on the market more than 50 years ago, the doses were high and people experienced a host of adverse reactions, including cardiac and blood pressure problems. Over the years, however, new birth control pills have required markedly lower doses.

"The expectation would be that the dose, at such a minimal level, will be effective in delaying the onset of mild cognitive impairment while alleviating many of the side effects" seen with thiazolidinediones, Roses added. "We will know that when we have seen the data."