The lymphoma drug bexarotene, under cross-examination as an Alzheimer’s therapeutic, dramatically lowers soluble Aβ in mice in some studies but not others. A paper in the September 12 Journal of Biological Chemistry offers an explanation for the discrepancy. Researchers led by Mary Jo LaDu at the University of Illinois at Chicago reported that bexarotene squelched soluble Aβ only in mice that expressed human apolipoprotein E4 (ApoE4), a major Alzheimer’s risk factor, and only in brain regions with extensive amyloid deposits. The authors attribute this to the drug’s ability to stimulate the packing of lipids onto ApoE4. This variant normally hooks less lipid than the other isoforms do, and may clear Aβ ineffectively as a result. In mice carrying human ApoE3, or in brain regions with little amyloid accumulation, bexarotene treatment provided no benefit.

“For ApoE4 carriers who have high Aβ levels, increasing ApoE lipidation may be a promising therapeutic strategy,” first author Leon Tai told Alzforum. Such a strategy might help people at later stages of AD, in contrast to current approaches that focus on prevention, Tai noted.

“The findings are very important,” said Ling Li at the University of Minnesota, Minneapolis. Other commentators agreed. “This is a promising lead that suggests drugs focused on ApoE could be an alternative to Aβ therapies,” said Michael Vitek, founder and CEO of biotech startup Cognosci Inc. in Research Triangle Park, North Carolina, which develops ApoE mimetics.

Bexarotene first attracted the attention of AD researchers when Gary Landreth and colleagues at Case Western Reserve University, Cleveland, Ohio, reported that it rapidly cleared Aβ and improved cognition in AD mice (see Feb 2012 news story). Attempts to replicate the finding met with mixed results, however, with some groups reporting no treatment benefit (see May 2013 news story). A recent study from Daniel Michaelson and colleagues at Tel Aviv University in Israel found that the drug boosted ApoE lipidation in E4 knock-in mice, restoring their cognitive performance and synaptic markers to levels seen in E3 carriers (see May 2014 news story). 

LaDu and colleagues wanted to investigate bexarotene’s effect in mice that expressed both human amyloid precursor protein (APP) and ApoE variants. They used transgenic mice, generated by introducing human ApoE isoforms into the 5xFAD model, which develops aggressive amyloid pathology (see Jul 2011 conference story). With Gregory Thatcher, also at the University of Illinois at Chicago, the authors dissolved bexarotene in sesame oil and administered it either by gavage or in the animals’ water supply. Drug formulation had been a sticking point in the interpretation of prior mouse data, but Thatcher’s pharmacokinetic analyses revealed that in either form, bexarotene entered the brain and engaged its targets, the lipidating proteins ABCA1 and ABCG1.

In 6-month-old ApoE4-FAD animals, bexarotene given for a week halved the level of soluble, oligomeric Aβ in the hippocampus, as detected by commercial ELISAs that use an antibody said to be specific for the oligomeric form. Treatment also pumped up lipidation of ApoE4 and increased the amount of Aβ complexed with the lipoprotein, suggesting that ApoE4 could more effectively clear the peptide. In contrast, in 6-month-old ApoE3-5xFAD mice, and in the cortex of E4-5xFADs where Aβ levels were low, the drug did not affect ApoE. Lower Aβ levels in these regions associated with higher ABCA1 and ABCG1 and greater ApoE3 and ApoE4 lipidation. The data implies that as Aβ rises, ApoE lipidation falls, and bexarotene counteracts that. There may be some threshold of ApoE lipidation, above which bexarotene helps no more, the authors suggest.

Surprisingly, for 6-month-old ApoE3 and ApoE4 FAD mice, one week of bexarotene treatment more than doubled the amount of soluble oligomeric Aβ in the cortex. How this detrimental effect comes about is unclear. However, the finding may help explain a discrepancy, LaDu noted. Earlier studies used models in which Aβ levels varied widely from mouse to mouse, and included females, which accumulate more Aβ than males. Therefore, bexarotene’s efficacy may have fluctuated greatly among the animals, canceling out effects in some studies. By contrast, Tai’s study used only male mice from an aggressive model with more uniform amyloid levels, he said.

In a one-month prevention study in younger animals, the authors saw no benefit from bexarotene. When treatment started at five months, neither ApoE lipidation nor soluble Aβ budged in either mouse model or brain region. Bexarotene is known to damage the liver, which roughly doubled in size in these animals. Since other studies have shown that liver toxicity can stress the central nervous system, perhaps this underlies the drug’s lack of efficacy during longer-term treatment, the authors speculated. “We think this indicates that systemic, off-target effects might be doing something harmful,” Tai said.

Bexarotene binds and activates nuclear transcription factors called retinoid X receptors, but also affects other nuclear receptors. The bexarotene analogue LG100268 binds more selectively to RXRs. Tai and colleagues tested it as well, but obtained identical results to bexarotene, suggesting that any RXR agonist might share similar problems. “We need to address these negative side effects of RXR agonists before we proceed to a clinical trial,” Tai said.

To other researchers, this study implies that the ApoE4 allele represents a loss of function. If so, raising ApoE levels or otherwise restoring its activity would be therapeutic. The field still debates this issue, as other work suggests lowering ApoE as the better strategy (see Apr 2013 news story). 

The paper raises new questions, Li noted. Is the disparity between bexarotene’s effects in the hippocampus and cortex due to amyloid pathology, or regional brain differences? ABCA1 and ABCG1 may be regulated differently in distinct regions, she noted. Does bexaroten affect other lipoproteins, such as ApoA1 (see Oct 2010 news story) or ApoJ (clusterin, currently No. 3 on AlzGene)? It would be interesting to directly measure Aβ clearance through the interstitial fluid and look at bexarotene’s effects on behavior in these mice, Li suggested.

Landreth said that the findings largely agree with previous work on bexarotene. “This is an interesting and provocative study that sheds new light on the complexities of drug action in the AD brain,” he told Alzforum. He noted that bexarotene is less toxic in people than in mouse models, and is given at lower doses. Liver enlargement occurs transiently in people, and the Food and Drug Administration has not recorded any central nervous system side effects in the tens of thousands of people who have taken this drug to combat skin cancer, he said. Landreth is wrapping up sample collection in a Phase 1b trial that examines the effect of the drug on ApoE and Aβ levels over 72 hours in healthy volunteers. A Phase 2 study of bexarotene in AD patients is ongoing at the Cleveland Clinic Lou Ruvo Center for Brain Health in Las Vegas.—Madolyn Bowman Rogers


  1. The pivotal publication from the Landreth group that the RXR agonist and anti-cancer drug bexarotene lowers the levels of Aβ in vivo and that this is associated with increased levels of ApoE (Cramer et al., 2012) paved the way for a new therapeutic approach to AD and simultaneously raised important questions. This paper from the LaDu group addresses these issues and extends them in an important way.

    The original Landreth study was performed with mice that express mouse ApoE (mApoE) and human Aβ (hAβ). However, the situation becomes theoretically more complicated in the case of human ApoE, because bexarotene could theoretically have opposing effects. Namely, upregulation of the expression of human ApoE, including ApoE4, which is the most prevalent risk factor of AD, could induce pathology, whereas upregulation of the ApoE lipidating protein ABCA1, which is also regulated by the RXR system, could theoretically be beneficial. Anat Bohem-Cagan from our group addressed this question recently utilizing mice that express human ApoE4 and ApoE3 and mouse Aβ. This revealed that bexarotene reverses key brain pathological and cognitive impairments that are induced by ApoE4 and that biochemically this is associated with reversal of the hypolipidation of ApoE4 and with no changes in ApoE4 levels (Boehm–Cagan et al., 2014). One of the ApoE4-driven phenotypes that was reversed by bexarotene in our study was the accumulation of Aβ in hippocampal neurons; however, this was performed on mice expressing only endogenous Aβ.

    In the present study from Tai and colleagues in the LaDu group, the authors developed and utilized a new, powerful, double-transgenic mouse system, which expresses human APP/Aβ and human ApoE3 or ApoE4, to study the effects of bexarotene on a background of human Aβ and human ApoE. They revealed that bexarotene increases the levels of ABCA1 and the lipidation of ApoE in these animals, and demonstrated that the latter results in an increased level of ApoE4/Aβ complex, which in turn leads to decreased levels of oligomeric Aβ. Since oligomeric Aβ is a key pathological player, this provides a novel mechanism whereby the pathological effects of ApoE4 are counteracted by bexarotene. Interestingly, the study also showed that the effects of bexarotene are brain-area-specific and are more pronounced in the hippocampus than the cortex.

    Importantly, bexarotene was efficacious when administered for seven days, but had no effect when over 30 days. Further studies are required to unravel the mechanisms underlying this effect and for the development of efficacious chronic treatments with bexarotene.


    . Reversal of apoE4-driven brain pathology and behavioral deficits by bexarotene. J Neurosci. 2014 May 21;34(21):7293-301. PubMed.

    . ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models. Science. 2012 Mar 23;335(6075):1503-6. Epub 2012 Feb 9 PubMed.

    View all comments by Daniel Michaelson
  2. In this study, Tai et al. explore the effect of two RXR agonists (bexarotene and LG100268) on amyloid pathology of APP transgenic mice (5XFAD) expressing human ApoE isoforms. Authors used short (seven days) and long (30 days) treatments to test the effect on Aβ and ApoE level. As a result of the short treatment they found an increase in ApoE protein level and a decrease of soluble Aβ and Aβ oligomers in a TBS-X-extracted fraction of the hippocampus of E4FAD mice. Unexpectedly, the opposite effects were observed in the cortex of E4FAD and E3FAD mice. Notably, in this AD model the major anatomical region of Aβ accumulation is the hippocampus, which incidentally has a higher ApoE protein level. Surprisingly, no effect on Aβ or other markers were found in the hippocampus of E3FAD mice. Since ApoE level in E4FAD is lower than in E3FAD mice, the authors argue that RXR agonists may elicit beneficial effects only when lipidated ApoE falls below some critical pathological threshold, as for example in E4FAD mice. The conclusion is that in the hippocampus RXR agonists increase ApoE lipidation and the formation of ApoE-Aβ complexes, which in turn decrease Aβ in this region. Such a hypothesis is supported by the finding that there is a negative association between ApoE-Aβ complex and oligomeric Aβ in the hippocampus of E4FAD mice.

    The study is interesting in that it helps to understand the effect of RXR agonists on amyloid pathology. However, to some degree it adds to the controversy surrounding the effects of bexarotene. It is important that the authors show an effect on amyloid pathology in AD mice expressing human ApoE4 isoforms, since ApoE4 carriers usually fail to respond to therapy in the ongoing clinical trials. In general the data also agree with published studies showing Bexarotene decreases soluble (Veeraraghavalu et al., 2013; Cramer et al., 2012; Fitz et al., 2013; Boehm-Cagan and Michaelson, 2014) but not insoluble Aβ levels in APP tg mice (Veeraraghavalu et al., 2013; Fitz et al., 2013; Tesseur et al., 2013; Price  et al., 2013). Interestingly, Tai et al. confirm our previous finding that Bexarotene decreases the level of soluble Aβ oligomers in the brain of APP mice expressing human APOE3 and E4 isoforms (Fitz et al., 2013). The authors also emphasize the significance of ApoE-Aβ complexes, which was recently questioned by Verghese et al. (Verghese et al., 2013). Finally, as in many studies, there are unanswered questions here as well: What was the effect of RXR ligands on amyloid plaques as detected by immunohistochemistry? Did Bexarotene affect cognitive performance/memory deficits in this AD model and how?


    . 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.

    . ApoE-directed therapeutics rapidly clear β-amyloid and reverse deficits in AD mouse models. Science. 2012 Mar 23;335(6075):1503-6. Epub 2012 Feb 9 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.

    . Reversal of apoE4-driven brain pathology and behavioral deficits by bexarotene. J Neurosci. 2014 May 21;34(21):7293-301. 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-d. PubMed.

    . ApoE influences amyloid-β (Aβ) clearance despite minimal apoE/Aβ association in physiological conditions. Proc Natl Acad Sci U S A. 2013 May 7;110(19):E1807-16. PubMed.

    View all comments by Radosveta Koldamova

Make a Comment

To make a comment you must login or register.


News Citations

  1. Upping Brain ApoE, Drug Treats Alzheimer's Mice
  2. Bexarotene Revisited: Improves Mouse Memory But No Effect on Plaques
  3. Has ApoE’s Time Come as a Therapeutic Target?
  4. Chicago: 2nd Annual Symposium Explores ApoE
  5. ApoE Does Not Bind Aβ, Competes for Clearance
  6. ApoA1: Does Good Cholesterol Protect the Brain?

Research Models Citations

  1. 5xFAD (B6SJL)

External Citations

  1. ApoJ
  2. Phase 1b trial
  3. Phase 2 study

Further Reading

Primary Papers

  1. . Amyloid-β pathology and APOE genotype modulate retinoid X receptor agonist activity in vivo. J Biol Chem. 2014 Oct 31;289(44):30538-55. Epub 2014 Sep 12 PubMed.