The cancer drug bexarotene grabbed headlines in 2012, when researchers led by Gary Landreth at Case Western Reserve University, Cleveland, Ohio, reported in Science that the drug swiftly cleared amyloid-β and restored memory in transgenic mouse models of Alzheimer's disease (see ARF related news story). That only two weeks of treatment nearly abolished existing amyloid plaques excited many researchers and catapulted the story into the popular press (see, e.g., Forbes story; CNN story; ABC News story). Curiously, longer-term treatment failed to budge these brain deposits, though cognitive improvements persisted. Now, in technical comments published in tomorrow's Science, four independent groups report that, in their hands, plaques stayed put in transgenic mice given bexarotene for one to two weeks, contradicting Landreth’s findings. However, several of the studies confirmed memory improvements and drops in soluble Aβ in treated mice. Bart de Strooper at KU Leuven, Belgium, senior author on one of the papers, told Alzforum, “We are far from understanding how this drug works. More research is needed before moving forward.”

Landreth finds the results encouraging. “The data provided in the technical comments replicate and validate the principal conclusions of our work,” he wrote in his published response. The findings strengthen the hypothesis that the drug works by lowering soluble Aβ peptide, not plaques, he noted. Landreth and colleagues have begun a Phase 1b clinical trial to examine bexarotene’s effect on Aβ production and clearance in healthy young volunteers.

Researchers who were not involved in the studies said the data were mixed. “If you put everything together, there seems to be a signal in terms of behavioral benefit and reducing soluble Aβ, although rapid plaque clearance is not supported,” said Dave Morgan at the University of South Florida, Tampa. Paul Aisen at the University of California, San Diego, said that bexarotene still looks like a promising drug, although the failure to replicate the dramatic effect on plaques “does dampen enthusiasm.”

Bexarotene, marketed as Targretin®, stimulates the transcription of apolipoprotein E and is approved by the U.S. Food and Drug Administration for the treatment of certain skin cancers. In Landreth’s original study, APP/PS1 mice taking the drug in their chow had a 25 to 30 percent drop in soluble Aβ in the interstitial fluid (ISF) within hours and fewer plaques within days, reaching up to 80 percent clearance after two weeks. Treatment improved several measures of cognition in various mouse models, Landreth reported.

The results were so eye-catching that many researchers quickly attempted to replicate them. In addition to de Strooper’s team, a second group was led by Kevin Felsenstein, Todd Golde, and David Borchelt at the University of Florida, Gainesville, and a third by Radosveta Koldamova and Iliya Lefterov at the University of Pittsburgh, Pennsylvania. The fourth group, led by Sangram Sisodia at the University of Chicago, Illinois, included numerous prominent AD researchers at different institutions, among them David Holtzman at Washington University School of Medicine, St. Louis, Missouri; Mathias Jucker at the University of Tübingen, Germany; Rudy Tanzi at Massachusetts General Hospital, Charlestown; and Robert Vassar at Northwestern University, Chicago. Informal discussions among the authors led to coordinated submission to Science.

Each group used somewhat different methods, mouse models, and drug formulations, but none saw any movement in plaques. Only the Belgium and Pennsylvania teams tested behavior; both saw recovery of memory to wild-type levels after two weeks of treatment. Soluble Aβ measurements proved the most variable. The Belgium and Florida groups found no change, while the Chicago group reported significant drops in soluble Aβ40 and Aβ42 in the 5XFAD mouse, but not in two strains of APP/PS1 mice used by Landreth. The Pennsylvania group was the only one to measure ISF, and reported reductions in Aβ similar to those seen by Landreth. They also found drops in soluble oligomers of Aβ, as visualized by antibodies specific for that form. Overall, the findings match data from some human studies, where plaques correlate poorly with cognitive change, Koldamova said. Likewise, Morgan noted that many antibody studies in mice have shown improved memory without a change in plaque load.

Landreth pointed out that the formulation of bexarotene may be a critical factor in the drug’s effects and could explain some discrepancies among the groups. The FDA-approved form consists of microcrystals of the drug suspended in an aqueous solution that keeps them stable, Landreth said (see FDA filing). Only the Pennsylvania group used this formulation. The other three used various organic solvents to dissolve bexarotene, in some cases starting with the free acid form of the drug instead of microcrystals. These preparations may change the bioavailability and pharmacokinetics of the drug, Landreth suggested. While all four groups demonstrated upregulation of bexarotene targets in the brain, most did not do pharmacokinetic studies. The Belgium group did, and reported 20-fold greater drug levels in brain than those used by Landreth, along with signs of toxicity such as weight loss, irritability, and difficulty breathing.

Issues with reproducibility have plagued the AD field and research science in general (see ARF related news story). Aisen advocates documenting preclinical study designs in a formal registry, similar to the way clinical trials are listed. This would include spelling out detailed methods, outcomes, and statistical analysis plans to facilitate future replication and better predict clinical trial results. De Strooper believes that the AD field would benefit tremendously if more follow-up studies, whether negative or positive, were published. “Journals such as Nature and Science that publish high-profile articles have a responsibility to provide a forum for follow-up work,” he said. The researchers praised Science’s decision to publish these comments. “This is a good model for moving forward. We should continue to pursue rapid attempts at replication of major findings,” Aisen said.

The authors of several of the comments believe that more research is needed before bexarotene goes to clinical trials. “We should proceed with caution,” Felsenstein said. He suggested replicating the results in non-human primates or dogs before moving to people. He also noted that there may be other compounds that have the same mechanism of action as bexarotene but with a better safety profile. Bexarotene produces a range of negative reactions, including high cholesterol, hypothyroidism, skin irritation, and sensitivity to light. In an e-mail to Alzforum, Joachim Herz at the University of Texas Southwestern, Dallas, suggested that this class of drugs may never be useful for chronic treatment in people due to the numerous problematic side effects.

Landreth’s Phase 1b study, supported by the BrightFocus Foundation and the Alzheimer’s Drug Discovery Foundation (see press release), will enroll 12 healthy young adults, half of whom will receive a single dose of the drug, and half placebo. In collaboration with St. Louis-based C2N Diagnostics, the researchers will measure changes in the synthesis and clearance of ApoE and Aβ in blood and cerebrospinal fluid (CSF) over 72 hours, using a method pioneered by Holtzman and Randy Bateman at WashU (see ARF related news story), co-founders of the company. “It’s a proof-of-mechanism trial. We feel this is a prerequisite to any subsequent development,” Landreth said. He expects the trial to wrap up by the end of the year.

Aisen noted that such a trial is an appropriate next step. “Pursuing a CSF clearance study early makes sense. It’s much more straightforward than the typical development pathway for potential AD-modifying drugs.” also lists a Phase 2 study of bexarotene in AD patients run by the Cleveland Clinic Lou Ruvo Center for Brain Health in Las Vegas, Nevada. Landreth is not involved in that trial.––Madolyn Bowman Rogers


  1. In our comments to Alzforum on 17 February 2012, we reported our failure to confirm the data of Cramer et al. in the APPSwFILon, PSEN1*M146L*L286V (5XFAD) mouse model of Alzheimer's disease. Mice were fed bexarotene 100 mg/kg/day for up to 50 days. We found no difference between test and control mice in a standard Morris water maze test, and no significant difference in the number of Aβ deposits. We did not publish these negative results, and only suggested that further in-depth investigation was indicated. That has now taken place with additional failures to confirm using other Tg mouse models. Given such data, it is difficult to be optimistic that bexarotene will be effective in treating AD.

  2. Phase 2A Trial of Bexarotene in Patients with Alzheimer’s Disease
    Cramer and colleagues (Cramer et al., 2012) reported the rapid reduction of soluble amyloid, as well as amyloid plaques, in two species of transgenic mice. The effect occurred over a period of only a few days and correlated with improved performance on murine cognitive tests. This report elicited substantial excitement in the scientific community, as the pathway by which bexarotene exerts its effects through the retinoid X receptors (RXR) has been subject to relatively little study and might represent a new avenue of therapy for Alzheimer’s disease (AD). Attempts to reproduce this work (Fitz et al., 2013; Price et al., 2013; Tesseur et al., 2013; Veeraraghavalu et al., 2013) confirm some aspects of the original report and fail to replicate others. The two studies assessing memory (Fitz et al., 2013; Veeraraghavalu et al., 2013) both showed enhanced cognition following treatment with bexarotene. All studies showed target engagement with induction of ApoE isoforms and effects on ABCA1. Fitz and colleagues (Fitz et al., 2013) showed a remarkable effect on soluble Aβ, reducing the level of Aβ oligomers by 50 percent. None of the studies was able to reproduce the reported effect on fibrillar amyloid and plaque burden. Some studies cautioned against pursuing studies of bexarotene in AD patients (Tesseur et al., 2013; Veeraraghavalu et al., 2013). The Cleveland Clinic Lou Ruvo Center for Brain Health has initiated a Phase 2A study of bexarotene in patients with AD. Patients with mild to moderate cognitive impairment or AD who have been verified by the presence of an amyloid burden on amyloid imaging are entered into a double-blind, randomized, placebo-controlled trial. One dose of bexarotene will be studied. The double-blind portion of the trial is one month in duration. The primary outcome is the reduction of plaque burden on amyloid imaging, and secondary outcomes include a variety of other biomarkers and clinical outcomes, including novel computerized assessments. The implementation of a human clinical trial of bexarotene might be viewed as premature. Given the uniformly fatal nature of AD, the current absence of disease-modifying therapies, the preliminary evidence from animal models to AD, and the large available safety database indicate that a clinical trial of bexarotene in patients with AD is warranted. Preclinical studies provide preliminary information regarding construction of human clinical trials. The high rate of non-reproducibility of preclinical outcomes (Gunawardena) and the inconsistency exhibited by the available studies of bexarotene, even when conducted by expert experimentalists, are daunting. The variability of outcomes in these preclinical studies might function as a case study in the challenges of translating preclinical observations to clinical research. The utility of bexarotene for human AD can be resolved only by studying human AD. Studies of pharmacodynamics of the effects of bexarotene on ApoE in healthy controls will provide important biological information, but these effects might differ markedly from pharmacodynamics in AD where amyloid clearance, aggregation, and deposition are dramatically different from those in healthy volunteers. A notable advantage of repurposing drugs approved for other indications that exhibit pharmacodynamic features relative to the treatment of AD is the safety data available on these agents. Bexarotene has been used in the treatment of many elderly patients with cutaneous T cell lymphoma since its approval. Moreover, AD patients and their caregivers have shown a high tolerance for side effects if treatments have promise of ameliorating their disease (Bearer et al., 2007). Thorough monitoring of both known and unknown side effects is required for responsible human clinical trials, and many safety measures are built into the clinical trial being performed at the Cleveland Clinic Lou Ruvo Center for Brain Health. The clinical trial we are conducting will not solve the issue of whether bexarotene is a potential therapy for AD. It will not confirm or disprove any aspect of the amyloid hypothesis or the RXR pathway. It will resolve the issue of whether bexarotene given at a specific dose for a specific amount of time and measured with specific outcomes has measurable effects in patients with AD. This is the most information that any clinical trial can be expected to provide. A beneficial effect of bexarotene will be followed by further exploration of dose responses, the potential value of intermittent therapy to minimize toxicity, the testing of other agents with effects on RXR-related pathways, and optimization of a related molecule with a better adverse event profile. A negative outcome will be followed by a comprehensive review of dose and duration of therapy, optimal outcomes, and integration of emerging preclinical data. The clinical trial being implemented at the Cleveland Clinic Lou Ruvo Center for Brain Health is a logical next step for exploring this potential pathway of therapy for AD. The trial is given increased urgency by the failure to produce disease-modifying effects through other approaches.


    . ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models. Science. 2012 Mar 23;335(6075):1503-6. PubMed.

    . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-c. PubMed.

    . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-d. PubMed.

    . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-e. PubMed.

    . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-f. PubMed.

    . Believe it or not: how much can we rely on published data on potential drug targets?. Nat Rev Drug Discov. 2011 Sep;10(9):712. PubMed.

    . Older Americans' risk-benefit preferences for modifying the course of Alzheimer disease. Alzheimer Dis Assoc Disord. 2009 Jan-Mar;23(1):23-32. PubMed.

  3. Like many others, our group was excited about the findings published by Gary Landreth's group (see Cramer et al., 2013) and we repeated the main experiments with short-term dosing of bexarotene in the mutant APPSWE/PS1ΔE9 mouse model. While the ABCA1 target was upregulated in APP/PS1 mice treated with bexarotene, this drug failed to attenuate Aβ plaques or fear-conditioning deficits. In addition, we are unable to replicate a robust and persistent fear-conditioning deficit in APPSWE/PS1ΔE9 mice even with an alternate, highly sensitive testing paradigm. Instead, we found significant gender differences present in plaque load and the fear-conditioning task, both primary measures used by Cramer et al.

    Bexarotene has now been tested in several animal models of Aβ amyloidosis besides APPSWE/PS1ΔE9, including APP/PS1-21 and 5XFAD (Veeraraghavalu et al., 2013) and APPSWE/PS1ΔE9 with APOE3/4 allele knock-ins (Fitz et al., 2013). All have failed to replicate a change in soluble or plaque forms of Aβ in brain tissue, with the exception of a single positive finding from Verraraghavalu et al. that the non-pathogenic, soluble Aβ40 decreased after acute treatment with bexarotene in the 5XFAD mouse model, but not in the APPSWE/PS1ΔE9 or the APP/PS1-21 models. Fitz et al. reported a significant decrease in interstitial fluid Aβ levels, though this was not replicated in their ELISA analysis of brain tissue for soluble or insoluble Aβ.

    Together, the array of published work on the effect of bexarotene on pathology and cognitive impairment in mouse models of Aβ amyloidosis produces no rigorous evidence that bexarotene is a suitable candidate for the treatment of Alzheimer’s disease.


    . ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models. Science. 2012 Mar 23;335(6075):1503-6. PubMed.

    . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-f. PubMed.

    . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-d. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Upping Brain ApoE, Drug Treats Alzheimer's Mice
  2. Guidelines at Nature Aim to Stem Tide of Irreproducibility
  3. Keystone: Pulse-Chasing AD Biomarkers, Snaring γ-Secretase Targets

External Citations

  1. Forbes story
  2. CNN story
  3. ABC News story
  4. 5XFAD mouse
  5. FDA filing
  6. press release
  7. Phase 2 study

Further Reading

Primary Papers

  1. . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-f. PubMed.
  2. . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-e. PubMed.
  3. . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-d. PubMed.
  4. . Comment on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-c. PubMed.
  5. . Response to comments on "ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models". Science. 2013 May 24;340(6135):924-g. PubMed.