Although the ApoE4 allele represents the primary genetic risk factor for late-onset Alzheimer’s disease, few researchers have attempted to target it therapeutically. Exactly how the allele promotes AD is only gradually becoming clearer, and the literature conflicts on such basic questions as whether raising or lowering ApoE would be the way to go. Two new papers in the May 21 Journal of Neuroscience strengthen the case for targeting ApoE, although the authors employ quite different strategies. Researchers led by David Holtzman at Washington University in St. Louis report that an antibody directed against mouse ApoE can clear amyloid deposits in AD model mice, while improving brain connectivity and learning. Meanwhile, researchers led by Daniel Michaelson at Tel Aviv University in Israel used the cancer drug bexarotene to increase lipidation of human ApoE4 in transgenic mice. This treatment reversed many of the pathologies associated with ApoE4, suppressing Aβ and phosphorylated tau while sharpening cognition. This finding supports the idea that the state of ApoE, rather than its level, may be crucial in determining its effects.

“The two papers provide further evidence that ApoE plays central roles in AD pathogenesis and is a legitimate target for therapy,” Gary Landreth at Case Western Reserve University, Cleveland, told Alzforum.

ApoE comes in three varieties: the harmful ApoE4 allele, the common ApoE3 allele, and ApoE2, which protects against Alzheimer’s disease (see Nov 2013 news story). ApoE4 has been most prominently linked to poor clearance of Aβ and greater amyloid accumulation (see Jul 2013 news story), though other mechanisms exist as well. Genetic studies by Holtzman’s group and others have found that halving the amount of ApoE3 or ApoE4 slows amyloid buildup (see Dec 2011 news story). They further showed that giving young AD mice an antibody to endogenous ApoE prevented Aβ deposits from forming (see Kim et al., 2012). 

This raised the question of whether the same strategy would work in mice that already had amyloid deposits. To test this, first author Fan Liao injected the anti-ApoE antibody HJ6.3 into 7-month-old APP/PS1 mice once per week. At this age, the mice actively deposit amyloid and have plaques. After five months of treatment, plaque load nudged down by 20 percent in the cerebral cortex and 40 percent in the thalamus. Curiously, plaques in the hippocampus stayed put. The authors saw no change in soluble Aβ40 in the brain, but the peptide jumped up by 50 percent in plasma. This might indicate enhanced clearance from brain, they suggest. 

Amyloid deposits (green) shrink after 14 days of anti-ApoE antibody treatment (see arrow, bottom panel). Courtesy, with permission of Liao et al., The Journal of Neuroscience, 2014.

To take a closer look, the authors followed the fate of individual plaques after exposing the surface of the brain to the HJ6.3 antibody for two weeks. Those plaques that contained ApoE and bound the antibody shrank or even disappeared over the course of treatment (see image). While there were fewer plaques, the overall effect on amyloid deposits was modest, as plaque load per square millimeter did not change, the authors noted. This contrasts with the antibody’s strong effect on preventing plaque formation in younger mice, Holtzman said.

It is unclear whether the antibody directly breaks up plaques by binding to ApoE, or stimulates microglia to gobble up plaques, Holtzman said. Intriguingly, the treatment only slightly lowered ApoE levels in the cerebral cortex. Soluble ApoE in the brain did not change, while the amount extractable by detergent dropped by about a quarter. This fraction might represent ApoE that resides in cell membranes or binds to the extracellular matrix, Holtzman said. 

The authors then investigated the clinical benefits of treatment. Treated mice learned better than their untreated littermates in the Morris water maze, and also showed improved brain connectivity by functional connectivity optical intrinsic signal imaging. This technique is similar to functional MRI, except that blood flow is measured by shining different wavelengths of light into the brain and observing how they are absorbed or reflected by oxygen through detectors attached to the skull. In the treated mice, brain regions synchronized their activity better than they did in the untreated animals.

Commentators praised the study and said it supports further research into ApoE-based therapeutics. “The results from this study are very encouraging, and if they hold true, it would mean that only a slight reduction of ApoE in a specific ApoE pool is needed to get beneficial effects,” Henrietta Nielsen at the Mayo Clinic in Jacksonville, Florida, wrote to Alzforum (see full comment below).

This study, however, did not examine how human ApoE would behave. Holtzman plans to address this in future work by crossing transgenic mice carrying human ApoE3 or ApoE4 with AD model mice. 

In the other paper, Michaelson and colleagues further studied the effect of bexarotene on human ApoE variants. Landreth’s lab had reported that this cancer drug lowers soluble Aβ and improves cognition in mice, although other researchers could not replicate these results (see Feb 2012 news storyMay 2013 news story). Bexarotene stimulates retinoid X receptor (RXR) transcription factors, and in mouse models pumps up the expression of endogenous ApoE and two enzymes that load lipids onto proteins, ABCA1 and ABCG1 (see Feb 2012 conference story). Previous research indicates that human ApoE4 lacks lipids compared to ApoE3.

Michaelson and first author Anat Boehm-Cagan fed bexarotene for 10 days to transgenic mice in which endogenous ApoE had been replaced by either human ApoE3 or ApoE4. The treatment amplified expression of ABCA1 and ABCG1, as expected, but surprisingly, not that of ApoE. Michaelson noted that this may represent a difference between human and mouse protein, as other researchers have told him they found similar results. In treated human ApoE4 knock-ins, the amount of high-molecular weight ApoE increased to the levels seen in the ApoE3 knock-ins. This indicates a boost in ApoE4 lipidation, Michaelson said. Since ApoE4 levels were comparable to levels in untreated mice, this experiment allowed the authors to examine only the effects of increased lipidation. 

They report that treatment normalized many aspects of the E4 knock-ins’ physiology. Levels of Aβ42 and phosphorylated tau in the hippocampus both dropped to levels seen in ApoE3 knock-ins. Meanwhile, levels of the presynaptic marker vesicular glutamatergic transporter 1 recovered. Treated ApoE4 knock-ins did better in the Morris water maze and novel object recognition tests, matching the performance of the E3 animals.

“The findings confirm the effect of bexarotene on cognitive performance. The lowering of phosphorylated tau is new and interesting, and should be explored further,” Radosveta Koldamova at the University of Pittsburgh told Alzforum. She was not involved in the research.

The data support increasing ApoE4 lipidation as a therapeutic strategy, Michaelson said. He is collaborating with biotech company Artery Therapeutics, Inc., of San Francisco to develop ABCA1 and ABCG1 agonists. The company will present preliminary data on this at the Alzheimer’s Association International Conference in July in Copenhagen.

Commentators agreed that this approach looks promising. “These data provide compelling evidence that increasing lipoprotein lipidation in the central nervous system may reverse the loss of function associated with ApoE4,” wrote Mary Jo LaDu and Leon Tai at the University of Illinois at Chicago (see full comment below).—Madolyn Bowman Rogers.


  1. The report by Michaelson's lab provides further validation of the effects of the RXR agonist bexarotene in murine models related to AD. They have employed a mouse model in which the human ApoE genes are knocked into the murine locus and, unlike previous studies, only the murine Aβ species are linked to the deficits in these mice. Importantly, they find bexarotene-mediated improvement in memory and learning in two different behavioral tasks, consistent with outcomes reported by Cramer et al., Fitz et al., and Tesseur et al. They report a drug-induced reduction in Aβ42 levels, similar to those reported by us (Cramer et al., 2012).

    Significantly, they have explored the effects of bexarotene on neurons and find that RXR activation reduces phosphorylated tau detected by the antibody AT8. Moreover, they argue that the drug has effects on the presynaptic marker VGlut1, elevating its abundance in the ApoE4 mice. There are a few unanswered questions. It is not clear why Aβ42 should be preferentially cleared and not Aβ40. They show a bexarotene-mediated reduction in AT8 immunoreactivity, but it is entirely unclear how this effect on tau is achieved. This work is noteworthy as it addresses the effects of bexarotene in a model that does not develop plaques, providing new evidence that plaques are functionally irrelevant.

    The Holtzman paper is very intriguing and follows from their 2012 paper. The new work reinforces the view that ApoE contributes in significant ways to regulation of amyloid deposition and that manipulation of ApoE levels with immunotherapy, or direct antibody exposure, alters plaque dynamics. The data suggest that immunotherapy applied after disease onset may be of therapeutic utility—at least in mice. The best effects on plaque reduction were seen in the thalamus, not a typical area evaluated in AD, with smaller or no effects seen in the cortex and hippocampus, respectively. The paper contains a few odd findings, including an effect on motor endpoints. The analysis of functional connectivity is a high-tech flourish that augments the principal findings of the paper. This is a nicely performed study that takes advantage of the substantial technical resources available to the Holtzman and Hyman labs. The caliber of the work is excellent and provides direct support for ApoE-based therapeutics.


    . 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.

    . 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.

  2. Both these studies are addressing the ApoE problem by using different strategies to alter ApoE concentrations.

    Liao and colleagues used a monoclonal mouse anti-ApoE antibody trying to sequester ApoE levels in plaque-bearing APP/PS1 mice at the age of 7 months. Using intraperitoneal injection of the HJ6.3 antibody for 21 weeks, the authors describe decreased Aβ plaque load, increased plasma Aβ levels, improved spatial learning, and decreased microglial activation in the cerebral cortex. These findings are very interesting given the continuous debate about whether to increase or decrease ApoE levels to combat its pathogenic effects. Interestingly, there was no effect on plasma ApoE levels upon immunization; there was, however, a slight reduction in ApoE from Triton-X soluble fractions from brain tissue lysates. The authors also did not observe any difference in plasma cholesterol. These latter results are surprising, as theoretically the administration site (periphery) would promote peripheral levels of ApoE as the first target upon antibody injection. It would be very interesting to figure out the titer distribution of this antibody. How much of it actually reaches the brain? The results from this study are very encouraging and if they hold true, it would mean that only a slight reduction of ApoE in a specific ApoE pool is needed to get beneficial effects. 

    In the second study, Boehm-Cagan and Michaelson have investigated the effect of the highly debated bexarotene on mRNA and protein levels of ApoE and the lipid transporters ABCA1 and ABCG1 in ApoE3 and ApoE4 targeted replacement mice.  Contrary to earlier published in vitro data (Cramer et al., 2012) they describe no effect of bexarotene on mRNA and protein levels of ApoE.  They do, however, describe an increase in both ABCA1 and ABCG1. Importantly, they found that the lipidation of ApoE4, as determined by immunoblotting, increased upon bexarotene treatment. The authors describe reversal of cognitive deficits in ApoE4 mice, and a reduction in Aβ accumulation and AT8 tau phosphorylation in both ApoE3 and ApoE4 mice. 

    The benefits described in both these studies appear to happen without any major influence on ApoE protein concentrations. This is intriguing. From our own studies on AD patients and controls, we know that CSF levels of total ApoE don’t differ between APOE genotypes. Similarly, the distribution of individual ApoE isoforms is about 50-50 in CSF from APOE3/4 carriers (Martinez-Morillo et al., 2014). Together these various findings suggest that not the levels of ApoE per se are important for AD pathology but rather qualitative characteristics—as suggested by Boehm-Cagan and Michaelson. 

    Last, in their discussion Boehm-Cagan and Michaelson mention that it’s important to note that bexarotene has peripheral effects on, for instance, lipid levels. This point is crucial, as the peripheral mechanisms are too often excluded from the discussion. For example, we recently found that despite the unaltered levels of the ApoE4 isoform in CSF from APOE4 carriers, the same individuals did exhibit a specific decrease in ApoE4 levels in plasma. In heterozygotes, the levels of the non-ApoE4 isoform were unaltered (Martinez-Morillo et al., 2014), thus demonstrating that the observed decrease in total plasma ApoE in APOE4 carriers is attributed to a specific decrease in the ApoE4 isoform. 

    Importantly, there are other described effects of rexinoids that are rarely mentioned. A Nature publication back in 1997 described antidiabetic effects of RXR ligands, which functioned as insulin sensitizers decreasing fasting glucose levels and reducing both hyperglycemia and hyperinsulinemia in mouse models of diabetes and obesity (Mukherjee et al., 1997). In regard to the documented effects on insulin on cognition, the anti-diabetic effects of bexarotene need to be looked into further.


    . 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.

    . Total apolipoprotein E levels and specific isoform composition in cerebrospinal fluid and plasma from Alzheimer's disease patients and controls. Acta Neuropathol. 2014 May;127(5):633-43. Epub 2014 Mar 15 PubMed.

    . Sensitization of diabetic and obese mice to insulin by retinoid X receptor agonists. Nature. 1997 Mar 27;386(6623):407-10. PubMed.

  3. As APOE4 is basically the genetic risk factor for AD, it is heartening to see these two papers.

    The paper by Holtzman and collaborators is a characterization that leaves few questions unanswered. That plaque burden decreases with systemic delivery of an ApoE antibody is very interesting. Coupled with the biochemical extraction that shows decreases in soluble Aβ and Aβ in the plasma, immunochistochemistry for microglial activation, extensive behavioral analysis, and evidence that lipoprotein biogenesis is not significantly disrupted by the removal of ApoE, the news here is excellent.  The addition of the topical application to live mice and the subsequent reduction in plaques is very compelling and adds another dimension to the interpretation of the results.  It will be interesting to see what effect the human APOE genotypes have on this antibody therapy.

    An ongoing debate in the field is whether ApoE4 represents a toxic gain or loss of function, and how this distinction will impact the activity of APOE-directed AD therapeutics. Michaelson and collaborators address this issue through treating ApoE-targeted replacement mice, which express mouse Aβ and human ApoE3 or ApoE4, with the anti-cancer drug bexarotene. Bexarotene, through increasing ABCA1/G1 expression, reversed the lipidation deficiency associated with ApoE4, without any effect on total ApoE levels. Importantly, bexarotene lowered intraneuronal Aβ42 and hyperphosphorylated tau levels and reversed cognitive deficits only in ApoE4-TR mice. These data provide compelling evidence that increasing lipoprotein lipidation in the CNS may reverse the loss of function associated with ApoE4. To capitalize on these findings, further research on the effects of APOE, human Aβ and ApoE-lipidation on AD progression are sorely needed. 

  4. There has been a growing realization that we need to expand AD drug targets beyond amyloid. Therefore, at first glance the news of ApoE as a potential therapeutic target would be welcomed. As other commentators have pointed out, the Holtzman paper presents solid data, based on state-of-the-art technology, to establish that ApoE immune therapy hits all the right milestones in APP/PS1 mice—decreased amyloid plaque load, improved spatial learning in Morris water maze etc.—and could be a potential treatment for human AD. However, upon some reflection it becomes less clear why this approach should prove effective in AD patients when the past efforts to lower Aβ plaque load by passive immunotherapy—which worked brilliantly in mice—failed in clinical trials.

    Notwithstanding the beautiful science, the rationale for ApoE immunotherapy in humans remains weak because it is the E4 variant of ApoE and not ApoE protein itself that increases AD risk. Moreover, although Aβ is invoked, the precise mechanism by which ApoE4 increases AD risk remains unclear. Young adult ApoE4 carriers also exhibit altered connectivity of the medial temporal lobes (Dennis et al., 2010), increased seizure susceptibility (Briellmann et al., 2000), propensity for HIV dementia (Corder et al., 1998) and impaired recovery following traumatic brain injury (Alexander et al., 2007), all of which are independent of Aβ. The AD field must look beyond Aβ to find an effective disease modifying treatment.

    On a side note, the authors showed that mice treated with anti-apoE antibody exhibited improved performance in a Morris water maze test (which is strongly dependent on hippocampal synaptic plasticity) without any decrease in hippocampal amyloid plaque load. Should we believe that hippocampal plaques do not matter? 


    . Temporal lobe functional activity and connectivity in young adult APOE varepsilon4 carriers. Alzheimers Dement. 2010 Jul;6(4):303-11. PubMed.

    . APOE epsilon4 genotype is associated with an earlier onset of chronic temporal lobe epilepsy. Neurology. 2000 Aug 8;55(3):435-7. PubMed.

    . HIV-infected subjects with the E4 allele for APOE have excess dementia and peripheral neuropathy. Nat Med. 1998 Oct;4(10):1182-4. PubMed.

    . Apolipoprotein E4 allele presence and functional outcome after severe traumatic brain injury. J Neurotrauma. 2007 May;24(5):790-7. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Averting a Late-life Crisis: Midlife ApoE2 Clears Plaques in Mice
  2. A Genetic Approach to the ApoE4 Puzzle
  3. Lowering ApoE Brings Down Amyloid in Mice
  4. Upping Brain ApoE, Drug Treats Alzheimer's Mice
  5. Bexarotene Revisited: Improves Mouse Memory But No Effect on Plaques
  6. San Francisco: Tweaking Brain ApoE Reduces Aβ, Symptoms

Research Models Citations

  1. APPswe/PSEN1dE9 (line 85)

Paper Citations

  1. . Anti-apoE immunotherapy inhibits amyloid accumulation in a transgenic mouse model of Aβ amyloidosis. J Exp Med. 2012 Nov 19;209(12):2149-56. PubMed.

Further Reading

Primary Papers

  1. . Anti-ApoE antibody given after plaque onset decreases Aβ accumulation and improves brain function in a mouse model of Aβ amyloidosis. J Neurosci. 2014 May 21;34(21):7281-92. PubMed.
  2. . Reversal of apoE4-driven brain pathology and behavioral deficits by bexarotene. J Neurosci. 2014 May 21;34(21):7293-301. PubMed.