Move over, β amyloid. Here comes apolipoprotein E—target of a potential new Alzheimer’s disease therapy that looks promising in mice. In an online Science paper published yesterday and covered widely in the lay media (see, e.g., National Public Radio; BBC News; The Wall Street Journal), Gary Landreth of Case Western Reserve University, Cleveland, Ohio, and colleagues report that AD mouse models rapidly clear amyloid-β and regain cognitive function when given an approved cancer drug that enhances production of endogenous ApoE.

ApoE transcription is stimulated by retinoid X receptor (RXR) heterodimers formed with either peroxisome proliferator-activated receptor γ (PPARγ) or liver X receptors (LXRs). Prior work shows that PPARγ and LXR agonists improve Aβ pathology and symptoms in AD mice (see Donkin et al., 2010; ARF related news story on Jiang et al., 2008; Heneka and Landreth, 2007). In the Science report—the formal publication of data Landreth presented at an ApoE symposium in Chicago last summer (see ARF related conference story)—first author Paige Cramer and colleagues hit both RXR-PPARγ and RXR-LXR pathways with a single compound—the RXR agonist bexarotene. This is a drug approved by the U.S. Food and Drug Administration to treat certain kinds of skin cancer caused by T cells.

Landreth’s team gave the RXR agonist orally to APP/PS1 mice at two, six, or 11 months of age. Using in-vivo microdialysis—a method pioneered by coauthor John Cirrito at Washington University School of Medicine in St. Louis, Missouri—the researchers saw a dip in soluble Aβ levels in brain interstitial fluid (ISF) of two-month-old APP/PS1 mice as soon as six hours post-treatment. By 24 hours, ISF Aβ42 dropped by 25 percent and remained so for more than 70 hours. The compound spurred a similar Aβ decrease in wild-type mice, but had no influence on ISF Aβ in ApoE-deficient animals, suggesting that ApoE is required for the Aβ-lowering effect.

Brain Aβ deposition typically shows up in these APP/PS1 animals at around six months of age, and here there were benefits, too, with mice showing less soluble and insoluble Aβ by ELISA, and fewer plaques in thioflavin S-stained hippocampus and cortex. Moreover, longer treatment (20-90 days) improved contextual fear conditioning and Morris water maze performance in APP/PS1 mice as well as in a transgenic APP mouse with more aggressive pathology (APPPS1-21). When given to Tg2576 mice, the compound helped restore social behavior such as nest-building—an ability that fades progressively in this AD strain (Wesson and Wilson, 2011), and sense of smell, which also wanes in Tg2576 mice. Odor-guided behaviors rely on neural networks in the piriform cortex (PCX) circuit, and disruptions in this circuit have been linked to olfactory impairment in AD patients and Tg2576 mice (Wesson et al., 2011). In the current work, Landreth and colleagues found aberrant PCX network activity in vehicle-treated Tg2576 mice, which returned to normal with three days of bexarotene. The Aβ pathology and behavioral changes were all accompanied by increased expression of ApoE, and other RXR-inducible genes, as measured by Western blots.

The study demonstrates that “an FDA-approved drug reduces amyloid burden in a spectacularly rapid and robust fashion. I have not seen something this exciting since the amyloid vaccine papers,” noted Steve Estus of the University of Kentucky in Lexington. “The FDA approval is key because this indicates that the drug has an acceptable safety profile, at least for limited use in humans.”

Based on their findings in AD mice, Landreth and colleagues plan to launch a Phase 1b biomarker trial of bexarotene in normal adults, looking for falling Aβ and rising ApoE levels in cerebrospinal fluid after treatment. The trial should start in the next few months and be finished by summer, Landreth told ARF. His team will collaborate with St. Louis-based C2N Diagnostics, a company founded to commercialize a method Randy Bateman and other WashU researchers developed to monitor Aβ synthesis and clearance in real time in human CSF (see ARF related news story).

One concern with LXR/RXR agonists is that they not only drive up transcription of ApoE and its transporters (such as ATP-binding cassette transporter, aka ABCA1), but also upregulate genes associated with fatty acid synthesis. This off-target effect has halted clinical development of LXR-targeted compounds, noted Kelly Bales of Pfizer Global Research and Development, Groton, Connecticut. At a recent ApoE workshop held at the Gladstone Institute of Neurological Disease in San Francisco, Bales reported on her team’s studies with other small molecules that drive up brain ApoE levels in AD mice.

Meanwhile, researchers led by Vijayalakshmi Ravindranath at the National Brain Research Centre in Haryana, India, have gone au naturel—reducing Aβ pathology and behavioral deficits in AD transgenic mice with extracts from an Indian ginseng (Withania somnifera). Surprisingly, the compound worked by boosting expression of low-density lipoprotein receptor-related protein (LRP) in the liver. (Brain LRP binds ApoE and helps regulate Aβ clearance.) These effects—reported online January 30 in the Proceedings of the National Academy of Sciences USA (Sehgal et al., 2012)—suggest it may be possible to develop AD therapeutic compounds that promote Aβ clearance without actually entering the brain.—Esther Landhuis


  1. The issue of whether to increase or decrease ApoE levels for AD therapy is confounded by the multiple functions of ApoE. In a landmark paper, Landreth and coworkers clearly demonstrate that bexarotene treatment, an RXR agonist, increases endogenous mouse ApoE levels, enhancing soluble Aβ clearance within hours, in APP/Aβ transgenic mice. Further, in this significant paper, bexarotene rapidly reduced Aβ plaque burden by greater than 50 percent in 72 hours and reversed cognitive deficits, and thus represents a novel AD therapy. These data are in apparent contrast with studies in which Aβ deposition is greatest with mouse-ApoE compared to the lack of ApoE (ApoE-/-) in multiple APP/Aβ transgenic models (e.g., publications from Fagan/Holtzman and Bales/Paul). However, these same studies further demonstrate that amyloid burden from greatest to least is: mouse-ApoE, ApoE-/-, ApoE4. Thus, one perspective on this issue of whether a therapeutic approach should increase or decrease ApoE is to compare ApoE4 to ApoE-/-. While ApoE4 has a beneficial effect on Aβ deposition compared to ApoE-/-, ApoE4 has a negative effect on LTP and Aβ-induced neurotoxicity. Indeed, the magnitude of LTP in both the absence (1) and presence (2) of oligomeric Aβ is greater with ApoE-/- mice compared to ApoE4 targeted replacement mice ex vivo. Furthermore, neuronal survival after oligomeric Aβ treatment is greatest in neurons co-cultured with glia isolated from ApoE-/- mice compared to neurons with glia from ApoE4 targeted replacement mice (3). Thus, although raising ApoE levels may exert a beneficial function on Aβ deposition, ApoE4 may also represent a toxic gain of function. Therapeutics that target specific aspects of ApoE function, of which bexarotene should be considered, may prove to be the most beneficial approach for AD.


    . ApoE isoform affects LTP in human targeted replacement mice. Neuroreport. 2004 Dec 3;15(17):2655-8. PubMed.

    . ApoE isoform-specific effects on LTP: blockade by oligomeric amyloid-beta1-42. Neurobiol Dis. 2005 Feb;18(1):75-82. PubMed.

    . Abeta42 neurotoxicity in primary co-cultures: effect of apoE isoform and Abeta conformation. Neurobiol Aging. 2007 Aug;28(8):1139-47. PubMed.

  2. The data presented in this paper show the most rapid removal of Aβ of any systemically applied therapy. They hold great promise for testing the amyloid hypothesis, and possibly for application in secondary prevention trials of asymptomatic patients where anti-amyloid therapies are most likely to be effective. However, reasonable dosages need to be identified before large-scale human applications can be considered. The extreme doses used in cancer studies did indicate some severe reactions to the agent, albeit rare. Our center (Byrd Alzheimer Institute) will not be writing prescriptions at this time, but look forward to well-controlled clinical trial opportunities.

    View all comments by Dave Morgan
  3. Some LXR ligands interact directly with γ-secretase. The structure of bexarotene suggests that it, too, may have a direct interaction with the complex. Given the rapidity of the effects observed, it would be interesting to see data for this compound in a cell-free γ-secretase assay.

    Whilst bexarotene is indeed FDA approved for oncology, the safety profile is unlikely to translate to chronic administration in an elderly population—particularly not at equivalent doses to those used. This would indicate that more potent/bio-available analogues will be required. Understanding the exact properties of the drug with regard to amyloid reduction will be critical.


    . Cholesterol independent effect of LXR agonist TO-901317 on gamma-secretase. J Neurochem. 2007 May;101(4):929-36. PubMed.

  4. I think this is a very intriguing paper. Most therapeutic programs have targeted amyloid peptide production, or immune-mediated clearance of Aβ. This novel strategy targets ApoE-mediated clearance of Aβ through activation of the retinoid X receptors. Since ApoE is by far the strongest genetic risk factor in sporadic AD, it is reasonable to assume that ApoE-related therapeutics may be disease modifying. That idea is backed by strong experimental data from studies in animal models. And since bexarotene is already FDA approved for another indication, it may advance rapidly into human studies.

    So this is very exciting work indeed. However, mouse models have not reliably predicted therapeutic effects in humans. Caution is crucial until human studies confirm target engagement (enhanced amyloid peptide clearance). Trials have already shown that two monoclonal antibodies reduce brain amyloid load as seen by PET imaging, so those agents may be more likely to succeed. However, the activity of bexarotene should be straightforward to confirm by looking for changes in the cerebrospinal fluid ApoE and Aβ. If the mouse pharmacodynamic effect is confirmed, bexarotene could be an excellent candidate for studies in prodromal and preclinical AD.

    View all comments by Paul Aisen
  5. This is very exciting work. With the RXR agonist bexarotene, a drug approved by the FDA, three different AD mouse models clear amyloid-β and regain cognitive function. It is amazing that it works so fast and dramatically. However, none of these models has tau pathology. How does it affect tau pathology? Before, we also had some good candidates that reduced Aβ plaque and rescued cognitive impairment in animal models, but in clinical trials, they failed. So we look forward to the next clinical trial with bexarotene.

    Also, ApoE4 increases the risk of developing the late-onset form of AD, and a lot of studies have shown that ApoE4 is toxic and that more ApoE4 leads to more plaques. If bexarotene enhances the expression of ApoE, then AD patients with E4 alleles will have more toxic ApoE4. I wonder if bexarotene will be beneficial to those patients. It has also been reported that ApoE protein level decreases on bexarotene. If so, then bexarotene may be of benefit. The effect of bexarotene probably depends on which is more important, the decrease of ApoE or the toxicity of ApoE4.

    View all comments by Jianhua Shi
  6. Before we break out the champagne, let us be aware that mouse ApoE and human ApoE are not the same!


    . Important differences between human and mouse APOE gene promoters: limitation of mouse APOE model in studying Alzheimer's disease. J Neurochem. 2007 Nov;103(3):1237-57. PubMed.

    View all comments by John Wojdyla
  7. The report by Cramer et al. is potentially exciting because it offers a plausible explanation as to why ApoE4 should be a risk factor for AD, and more importantly suggests that bexarotene, an off-the-shelf retinoid X receptor agonist, may prove to be an effective treatment for the disease. Unfortunately, our results with bexarotene treatment in the APPSwFILon, PSEN1*M146L*L286V (5XFAD) mouse model of Alzheimer's disease failed to provide positive results. This 5XFAD model is one of the most aggressive of all AD models, with Aβ deposits appearing as early as eight weeks after birth. It may be that the high rate of Aβ deposition in this model overwhelms the ability of bexarotene to stimulate clearance. However, it does indicate that testing should be done in other AD transgenic models to gain a fuller picture of what is taking place.

    We started to feed the 5XFAD mice at seven to eight weeks of age with chow containing 500 mg/kg bexarotene, which corresponds to a consumption of 2 mg per day or an oral dose of 100 mg/kg/day (n = 8). Control 5XFAD mice were fed normal chow (n = 7). Mice were sacrificed at 21, 31, and 50 days after feeding commenced. 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.

    Transgenic mouse models of Alzheimer's disease vary widely in their properties. The model we studied may be an unsatisfactory representation of AD pathology. However, the results do illustrate that the phenomenon described by Cramer et al. is not entirely straightforward. More in-depth investigation is clearly indicated.

  8. Reply to comment by Jian-Ping Guo, P.L. McGeer, and Claudia Schwab

    We communicated the results of our study in advance of publication with these authors.

    It is important that oral administration of bexarotene employ the micronized form of the drug (Targretin™, Eisai, Inc.) due to the efficiency of absorption in the gut. The FDA filing documents a sixfold difference in efficacy compared to the powdered form of bexarotene.

    Dr. Guo et al. note that their model might be too robust to observe an effect of bexarotene. We tested the effect of bexarotene in a similarly aggressive model of AD, the APP/PS1-21 mice. A 20-day treatment resulted in a reduction of soluble Aβ and an approximately 35 percent decrease in plaque number. Importantly, we found improved behavior in both the Morris water maze and contextual fear conditioning. Thus, I don't think that this consideration explains their results.

    The correspondents noted that they did not find a change in plaque load in drug-treated mice. Indeed, we observed this, too. One of the surprising findings was that chronic drug treatment of APP/PS1 mice was associated with an immediate fall in plaque burden, which then returned to the levels observed in vehicle-treated mice. We found improved behavior in the same two behavioral assays after 90 days of treatment. One possible conclusion from this experiment is that plaques don't matter, and we are actively pursuing this issue.

    It should also be noted that in the periphery, there is robust induction of P450 enzymes with sustained bexarotene administration. It is possible that the dose of drug required to maintain microglia in an "alternatively activated" (M2) state, which is anti-inflammatory and conducive to plaque phagocytosis, is greater than that necessary to drive ApoE/HDL induction. The sustained drug treatment may result in efficient drug degradation, lowering the CNS and plasma levels of bexarotene. We are now testing this possibility.

    We couldn't agree more that further in-depth study is necessary, and are anxious to have our studies repeated in others' hands.

    View all comments by Gary Landreth

Make a Comment

To make a comment you must login or register.


News Citations

  1. ApoE’s Secret Revealed? Protein Promotes Aβ Degradation
  2. Chicago: 2nd Annual Symposium Explores ApoE
  3. Keystone: Pulse-Chasing AD Biomarkers, Snaring γ-Secretase Targets

Paper Citations

  1. . ATP-binding cassette transporter A1 mediates the beneficial effects of the liver X receptor agonist GW3965 on object recognition memory and amyloid burden in amyloid precursor protein/presenilin 1 mice. J Biol Chem. 2010 Oct 29;285(44):34144-54. PubMed.
  2. . ApoE promotes the proteolytic degradation of Abeta. Neuron. 2008 Jun 12;58(5):681-93. PubMed.
  3. . PPARs in the brain. Biochim Biophys Acta. 2007 Aug;1771(8):1031-45. PubMed.
  4. . Age and gene overexpression interact to abolish nesting behavior in Tg2576 amyloid precursor protein (APP) mice. Behav Brain Res. 2011 Jan 1;216(1):408-13. PubMed.
  5. . Sensory network dysfunction, behavioral impairments, and their reversibility in an Alzheimer's β-amyloidosis mouse model. J Neurosci. 2011 Nov 2;31(44):15962-71. PubMed.
  6. . Withania somnifera reverses Alzheimer's disease pathology by enhancing low-density lipoprotein receptor-related protein in liver. Proc Natl Acad Sci U S A. 2012 Feb 28;109(9):3510-5. Epub 2012 Jan 30 PubMed.

Other Citations

  1. APPPS1-21

External Citations

  1. National Public Radio
  2. BBC News
  3. The Wall Street Journal
  4. bexarotene
  5. APP/PS1 mice
  6. ApoE workshop

Further Reading

Primary Papers

  1. . 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.