Questions on the Value of ‘Off-the-Shelf’ Drugs for Alzheimer’s

May 23, 2013

New studies have raised questions about the effectiveness of two FDA-approved drugs that had shown promise as potential Alzheimer’s treatments.

One of the drugs is bexarotene, a skin cancer drug that was reported last year to reduce amyloid beta (Aβ) levels and reverse cognitive impairments in Alzheimer’s mice. The other is IVIG, a solution of mixed antibodies from healthy volunteers, which normally is used to treat immune deficiency or autoimmune conditions but had seemed to stop the progress of dementia in small trials in people with Alzheimer’s.

Results from larger clinical trials now suggest that IVIG has little or no effect against Alzheimer’s dementia. Three new studies in mice suggest that bexarotene too has little effect—but these studies are far less conclusive, and bexarotene’s testing as a potential Alzheimer’s therapy seems likely to continue.

“The most important observations [in our 2012 study] are being repeated, validating our point of view and supporting the testing of bexarotene in humans,” says Gary Landreth, whose laboratory at Case Western Reserve University School of Medicine produced the 2012 findings.


Bexarotene (brand-name Targretin) is FDA-approved to treat cutaneous T-cell lymphoma; it works by activating the retinoid X receptor (RXR) in certain types of cells. As a cancer drug, it is aimed at cells in the body. But unlike most drugs, it also gets easily into the brain, where its activation of RXRs on astrocytes and microglial cells is thought to have two potential anti-Alzheimer’s effects. First, it boosts astrocytes’ production of apolipoprotein-E (apo-E), a molecule that somehow helps regulate the levels of Aβ in the brain. Second, it apparently enhances microglial cells’ ability to gobble up Aβ aggregates.

(In the brains of people with Alzheimer’s, Aβ clumps into small soluble aggregates called oligomers, which are believed to contribute to the dysfunction of memory-related neurons; the disease also features large, insoluble aggregates that form visible “plaques.” There is evidence that keeping Aβ levels low throughout life can greatly reduce Alzheimer’s risk.)

Early last year, Landreth, first author Paige E. Cramer, then a graduate student, and their colleagues reported in Science that bexarotene had dramatic effects in transgenic mouse models of Alzheimer’s. The drug, they wrote, sharply reduced Aβ plaques within days, dropped soluble Aβ brain levels by one-third, and also greatly ameliorated cognitive deficits, according to several behavioral measures. (See “Off-the-shelf drug rapidly clears Alzheimer’s protein in mice.”)

As a result of these findings, Landreth and Cramer set up a company, ReXceptor Inc., and recently secured philanthropic funding for a short, “phase 1” trial of bexarotene in 12 healthy people.

Now three groups of researchers have reported, in “technical communications” to Science published online on May 23, 2013, that they could not reproduce the full results of the 2012 Cramer et al study. Most strikingly, they found no significant reduction of Aβ plaques. Some of the new findings did indicate an effect of the drug on soluble Aβ levels and cognitive measures, but not consistently.

“We saw a statistically significant decrease in soluble Aβ levels in only one mouse model,” says University of Chicago School of Medicine researcher Sangram S. Sisodia, who led one of the studies. “In two other mouse models, which were also used by Landreth and colleagues, we didn’t see any statistically significant reduction in soluble Aβ,” said Sisodia, also a member of the Dana Alliance for Brain Initiatives.

Another group emphasized the toxic effects they observed in treated dogs and mice, and concluded alarmingly: “Given the toxicity of bexarotene, our study clearly strongly cautions against testing this drug in [Alzheimer’s] patients at this time.”

Though strongly worded, these findings have their own weaknesses, as Landreth and his colleagues argued in a rebuttal paper. To begin with, most of the experiments performed in the three studies used a different bexarotene formulation—a mix of the active ingredient in a delivery “vehicle” of oils, alcohols and/or other solvents and additives—than the standard FDA-approved formulation used in Targretin capsules. Apparently as a result, the drug quickly reached toxic levels in the treated animals, making any conclusion about efficacy dubious. One group reported bloodstream levels in mice that were 20 times higher than a standard formulation should have produced. “They overdosed those animals, and they were sick,” says Landreth.

In experiments on one type of Alzheimer’s mouse model, Sisodia’s group did make use of a standard water-based formulation and showed no effect on plaques or soluble Aβ, but these experiments involved only three mice (per treatment and control group) for the analyses of soluble Aβ—a sample that is probably too small to justify strong conclusions.

Landreth also points out that the plaque-reductions reported in Cramer et al were not dramatic in all their experiments, and that plaque reduction isn’t how the drug would work anyway. Soluble oligomers, not plaques, are now thought to be the principal harmful species of Aβ—the plaques are considered relatively non-toxic, and might even be protective to the extent that they sequester Aβ in non-oligomer form.

Alzheimer’s researcher Samuel Gandy at the Mount Sinai School of Medicine notes too that some transgenic Alzheimer’s mouse models are not reliable in how they manifest brain Aβ plaques, which might help explain the discrepancies in plaque-reduction results. “There is a lot of interindividual variability in plaque burden early on, when most people are looking for gene and drug effects,” Gandy says.

Perhaps most importantly, other groups have been finding signs that bexarotene might be useful against Alzheimer’s. Alongside the three adversarial reports this week in Science, a team at the University of Pittsburgh has described using bexarotene to treat transgenic Alzheimer’s mice that express human forms of apo-E. Although these researchers didn’t find significant plaque reduction, they did find a significant reduction of Aβ oligomer levels, and also a reversal of cognitive deficits. According to Landreth, two other laboratories, in Israel and the US, reported similar findings for bexarotene and a different RXR agonist at a recent Alzheimer’s conference. Moreover, last month the laboratory of David Holtzman at Washington University–St. Louis reported a new technique for evaluating apo-E-related experiments, and noted that in ordinary lab mice, bexarotene appears to boost apo-E levels in brain fluid while dropping Aβ levels. “So basically he directly validated the principal findings in our paper,” says Landreth.

“The RXR-apoE pathway may still be a good target for [Alzheimer’s drugs],” says Holtzman, also a Dana Alliance member. However, he cautions that “there are many basic questions that still need to be answered about this pathway.” Although RXR agonists may alter the levels or characteristics of apo-E, thereby affecting Aβ levels and related disease processes, they “also affect other pathways that we don’t yet understand.”

The debate has exposed some of the untidiness of science that normally is concealed or smoothed over in scientific publication. It may also inspire more in-depth research on RXR and apo-E related pathways in Alzheimer’s. But Landreth says that it won’t stop his company’s phase 1 trial of bexarotene’s ability to boost apo-E and reduce Aβ levels—a study which he expects to be completed and evaluated by January 2014. He notes, too, that although bexarotene’s basic patent protections will expire within the next few years, making it a problematic investment for most pharma companies, it isn’t the only possible drug compound that can activate RXRs and boost apo-E.

“I know that big pharma have already pulled out their other RXR [agonist] compounds and are looking at them,” Landreth says.


The IVIG story seems to leave less room for optimism. The intravenously delivered solution of mixed antibodies—a pooled human blood-product sold by a number of companies globally—has been reported to arrest the course of dementia in people with Alzheimer’s, in small-scale studies over the past several years. Just as impressively, it has been said to prevent the brain shrinkage—as measured by MRI—that normally is seen as dementia worsens. (See “Alzheimer’s immunotherapies: reasons for hope.”)

IVIG’s mechanism of action against Alzheimer’s, if any, is not known. Researchers have speculated variously that it draws down Aβ levels, targets oligomers of the Alzheimer’s-associated protein tau, and/or reduces Alzheimer’s-related neural inflammation.

However, earlier this year Octapharma, a Swiss maker of IVIG, reported that its six-month trial in 56 people with probable Alzheimer’s had failed to detect the hoped-for improvements as measured by bloodstream Aβ, cognitive tests and MRI scans.

On May 8, Baxter Healthcare, a US-based maker of an IVIG product, announced that it  too had failed to find robust evidence of the antibodies’ effectiveness against Alzheimer’s—this time in a much larger and more definitive “phase 3” study of 390 patients with mild to moderate Alzheimer’s disease.

Other trials by other IVIG makers continue, and in the coming months researchers can be expected to sift through all the data in the hope of finding a hint of benefit in one or more subgroups of patients. But IVIG looks set to join the long list of prospective Alzheimer’s therapies that have raised hopes in small trials only to dash them in larger ones.

“Baxter will reconsider its current approach for its Alzheimer's program and will determine next steps after full data analyses,” the company stated in its press release. “The current Baxter studies of [IV]IG in mild to moderate Alzheimer's disease will be discontinued.”