Does ApoE4, the strongest genetic risk factor for sporadic Alzheimer’s, prevent synaptic pruning? Counterintuitively, that is what scientists conclude in the September 6 Proceedings of the National Academy of Sciences. The researchers, led by Ben Barres of Stanford University in Palo Alto, California, in collaboration with David Holtzman and colleagues at Washington University in St. Louis, report that the risk allele hinders phagocytic activity in astrocytes, preventing them from nibbling on synapses. How could that promote dementia? They also report that human ApoE4 leads to a glut of the complement protein C1q in the brains of transgenic mice. The researchers speculated that C1q decorates uncleared, deteriorating synapses, which could drive neuroinflammatory responses in the brain as the animals age. Researchers believe C1q tags synapses for destruction by microglia in AD and potentially other neurodegenerative diseases (see Nov 2015 conference news). 

“This work is highly exciting,” said Guojun Bu of the Mayo Clinic in Jacksonville, Florida. “It introduces a new dimension to the diverse functions of ApoE as a risk factor for AD,” he added. Bu was not involved in the study.

Strange Complement.

The complement protein C1q accumulates in the mouse hippocampus when ApoE4 is expressed (right) but not ApoE2 (left). [Courtesy of Chung et al., PNAS 2016.]

Apolipoprotein E exists in three different flavors—ApoE2, ApoE3, and ApoE4. While ApoE2 reduces AD risk, ApoE4 elevates it. Researchers have long sought to understand how the apolipoprotein, the only one in the brain, exerts its influence over AD. Previous findings indicate that ApoE affects the uptake of Aβ by astrocytes, and that ApoE4 somehow inhibits this process (see Holtzman et al., 2000Koistinaho et al., 2004Apr 2013 news). However, transgenic knock-in mice expressing human ApoE4 display behavioral and synaptic deficits independent of Aβ pathology (see Bour et al., 2008Dec 2009 news). Furthermore, Barres and colleagues had previously reported that astrocytes, which express most of the ApoE in the brain, play a key role in refining neural circuitry by pruning synapses (see Dec 2013 conference news). This prompted the researchers to investigate the relationship between ApoE genotype and this synapse-eating activity of astrocytes.

First author Won-Suk Chung—who has since moved from Barres' lab to the Korea Advanced Institute of Science and Technology in Daejeon—started by asking whether ApoE alleles could differentially influence the phagocytic activity of astrocytes. Chung and colleagues designed a cell culture system to address this. The researchers grew astrocytes from ApoE knockout mice, then added culture medium from astrocytes derived from ApoE2, E3, or E4 transgenic knock-in mice. This culture medium contained ApoE as well as any other factors the astrocytes were pumping out. Chung then added phagocytic prey—fluorescent synaptosomes—to the mix, and monitored phagocytosis. He found that cultures bathed in medium from ApoE2-expressing astrocytes engulfed the most synaptosomes, while cultures bathed in ApoE4-laced medium took up the least. The same was true when the researchers used synaptosomes labeled with pHrodo, a dye that emits fluorescence only after arrival in the acidic belly of the astrocyte (i.e., acidic organelles such as the lysosome).

To determine if it was the ApoE in the medium that influenced phagocytosis, the researchers next added pure lipidated ApoE to ApoE knockout astrocyte cultures. The ApoE effect held up, but only if the researchers also added an opsonin. Opsonins bridge the gap between phagocytic receptors on cells and the cargo they engulf. This result indicted that ApoE proteins do not directly alter astrocyte phagocytosis, but instead boost or inhibit phagocytic pathways with the help of other proteins.

To extend their findings to a relevant in vivo model, the researchers investigated how the different ApoE alleles affected synapse pruning during early postnatal development, a task that falls to astrocytes as well as microglia. To quantify this, they focused on commonly studied neural connections in the eye that undergo massive pruning during development, namely the extension of retinal ganglion cell (RGC) axons into presynaptic terminals in the dorsal lateral geniculate nucleus (dGLN). In various ApoE knock-in strains, the researchers injected fluorescent molecules into the eye to label the RGC axonal projections, then monitored fluorescent uptake into astrocytes gathered around the synaptic terminals in the dGLN. As with their cell culture experiments, they found that astrocytes in ApoE2 knock-in mice had gorged on the most synapses, while those in ApoE4 animals had taken up the least. The intermediate level of phagocytosis in ApoE3 knock-in animals was on par with that in ApoE wild-type or knock-out mice, suggesting that ApoE2 and ApoE4 alleles had dominant phagocytosis enhancing and inhibiting effects, respectively.

The researchers hypothesized that the reduced level of synaptic clearance in ApoE4 carriers would lead to an accumulation of nonfunctional or “senescent” synapses, which could interfere with neural circuitry, trigger neurodegeneration, and partly explain increased susceptibility to AD. The researchers had previously reported that senescent synapses build up with age, and that they are decorated with the complement protein C1q (see Dec 2007 newsStephan et al., 2013). Therefore, they relied on C1q levels as a proxy for the number of such synapses in the hippocampi of different ApoE knock-in mouse strains. In keeping with their hypothesis, the researchers found low levels of C1q in the hippocampi of nine-month-old ApoE2 mice compared to ApoE3 or ApoE4 mice. By 18 months, ApoE4 knock-in mice had significantly more hippocampal C1q compared to the other strains.

Barres explained that the presence of synaptic C1q alone is insufficient to trigger the complement cascade that ultimately beckons hungry microglia to dine on the synapses. Rather, he proposed that some type of second hit triggers the cascade. While microglial pruning of synapses is beneficial during development, studies have suggested that in the context of AD, this pruning kicks into overdrive and hastens neurodegeneration instead (see Apr 2016 news). The connection between astrocytic and microglial pruning functions is somewhat blurry, but the current findings help bring it into focus, Barres said. “This is the first paper to link ApoE allele to buildup of synaptic C1q—and thus to synapse degeneration by the complement cascade,” Barres said. “The pieces are starting to come together.”

“These data provide a new mechanism for the APOE4-related neurodegeneration and APOE2-related neuroprotection, namely phagocytic capacity,” commented Mary Jo LaDu, Ana Valencia Olvera, Deebika Balu, and Maria Evangelina Avila-Munoz of the University of Illinois in Chicago in a joint comment to Alzforum. However, the commentators pointed out that some of the results contradict previous findings suggesting that C1q promotes neuronal survival and protects against synapse loss. “Critical to these questions is the interactions between astrocytes and microglia,” they added (see full comment below).

Other unanswered questions include how ApoE influences astrocyte phagocytosis. Barres and colleagues hypothesized it might interfere with the binding of synapses to phagocytic receptors on the astrocyte surface—either enhancing binding in the case of ApoE2, or inhibiting it in the case of ApoE4. Bu thinks ApoE may be uniquely positioned to do this. “ApoE binds to a variety of cell surface receptors including transport receptors LRP1 and LDLR, as well as signaling receptors ApoER2/LRP8 and VLDLR,” he said. “Whether one or more of ApoE receptors or an alternative receptor involved in regulating phagocytosis mediate the ApoE isoform-dependent effects requires further investigation.”—Jessica Shugart


  1. This work is highly exciting. It introduces a new dimension to the diverse functions of ApoE as a risk factor for AD. Although the primary function of ApoE expressed in astrocytes is to transport lipids such as cholesterol, additional roles in neuronal signaling, vascular homeostasis, and inflammatory responses have been proposed, and it seems to regulate both the clearance and aggregation of amyloid-beta peptides in the pathogenesis of AD. Thus, the newly defined function of ApoE in regulating the phagocytic function of astrocytes in an isoform-dependent manner suggests that an even more complex network of functions might collectively contribute to the increased risk for AD conferred by APOE4.

    Another interesting dimension of this work is the emphasis of a phagocytic function of astrocytes, which is better known for microglia, the resident innate immune cells in the brain that also express ApoE. The data supporting a role for ApoE in modulating phagocytic efficiency appears to be solid; however, the underlying mechanism is not clear. ApoE binds to a variety of cell surface receptors, including transport receptors LRP1 and LDLR, as well as signaling receptors ApoER2/LRP8 and VLDLR. Whether one or more of these ApoE receptors or an alternative receptor mediates the ApoE isoform-dependent effects on phagocytosis requires further study. Further, the in vivo relevance of this work requires investigation using model systems exhibiting AD-related pathologies. The human relevance is even more difficulty to define but perhaps can be addressed by analyzing pathways relevant to the phagocytic function of astrocytes in human AD brains with different APOE genotypes. Overall, this is a highly interesting finding that warrants follow-up studies on both mechanisms and relevance.

  2. This is an intriguing publication that continues the authors’ premise that astrocytes play an active role in synaptic phagocytosis. In 2013, Chung and collaborators proposed a new role for astrocytes in the adult brain, specifically phagocytosis activated not via the compliment cascade but through the MEGF10 and MERTK pathways (Chung et al., 2013). Stevens and co-workers demonstrated that factors secreted by activated astrocytes induce C1q secretion by neurons and microglia, initiating the complement cascade in microglia, leading to phagocytosis of synapses (Stevens et al., 2007; Stephan et al., 2012). C1q is of particular interest as it is upregulated during aging and neurodegeneration (Stephan et al., 2013) and has recently been demonstrated to mediate the early synapse loss in AD mouse models (Hong et al., 2016). This present study by Chung and collaborators incorporates the role of APOE genotype in these novel astrocytic pathways. The authors analyzed primary astrocytes from APOE-KI mice and developing brains to demonstrate that APOE modulates the capacity of astrocytes to engulf synaptosomes, in the order APOE2>APOE3>APOE4. In addition, age-induced C1q accumulation is reduced with APOE2 and increased with APOE4. These data provide a new mechanism for the APOE4-related neurodegeneration and APOE2-related neuroprotection, namely phagocytic capacity.

    As the authors point out, future studies are required to elucidate how APOE genotype regulates astrocytic phagocytosis, including interactions with the MEGF10, MERTK, and even complement pathways, and how this regulation is affected by APOE genotype, aging, and neurodegeneration. The interaction with C1q is also important but apparently contradictory as an increase in C1q has been reported to promote neuronal survival (Heese  et al., 1998; Pisalyaput and Tenner, 2008; Benoit  and Tenner, 2011), while C1q-knock out leads to a decrease in phagocytic microglia and early synapse loss (Stephan et al., 2012; Hong et al., 2016). Critical to these questions are the interactions between astrocytes and microglia. As well, in a long-term (28 days) neuron/glial co-culture model, apoE4 delays spine formation and accelerates the loss of mature spines, while apoE2 enhances both spine formation and maintenance (Nwabuisi-Heath et al., 2013). 

    To incorporate these findings with APOE associated AD risk, it is necessary to evaluate these pathways in an AD/APOE relevant model. Unpublished RNA-Seq data in our EFAD mice (human APOE expressed in 5XFAD mice), demonstrate that complement cascade pathways were significant in E3FAD vs E4FAD (three of the top four pathways with Metacore analysis), supporting the importance of these mechanisms in AD pathology. 


    . Astrocytes mediate synapse elimination through MEGF10 and MERTK pathways. Nature. 2013 Dec 19;504(7480):394-400. Epub 2013 Nov 24 PubMed.

    . The classical complement cascade mediates CNS synapse elimination. Cell. 2007 Dec 14;131(6):1164-78. PubMed.

    . The complement system: an unexpected role in synaptic pruning during development and disease. Annu Rev Neurosci. 2012;35:369-89. PubMed.

    . A Dramatic Increase of C1q Protein in the CNS during Normal Aging. J Neurosci. 2013 Aug 14;33(33):13460-74. PubMed.

    . Complement and microglia mediate early synapse loss in Alzheimer mouse models. Science. 2016 May 6;352(6286):712-6. Epub 2016 Mar 31 PubMed.

    . Complement and microglia mediate early synapse loss in Alzheimer mouse models. Science. 2016 May 6;352(6286):712-6. Epub 2016 Mar 31 PubMed.

    . Complement component C1q inhibits beta-amyloid- and serum amyloid P-induced neurotoxicity via caspase- and calpain-independent mechanisms. J Neurochem. 2008 Feb;104(3):696-707. PubMed.

    . Complement protein C1q-mediated neuroprotection is correlated with regulation of neuronal gene and microRNA expression. J Neurosci. 2011 Mar 2;31(9):3459-69. PubMed.

    . ApoE4 delays dendritic spine formation during neuron development and accelerates loss of mature spines in vitro. ASN Neuro. 2014 Jan 13;6(1):e00134. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Microglia Control Synapse Number in Multiple Disease States
  2. ApoE Does Not Bind Aβ, Competes for Clearance
  3. Gift of the GABA? Transmitter Nixes Neurogenesis in APP, ApoE4 Mice
  4. Glial Cells Refine Neural Circuits
  5. Paper Alert: Does the Complement Devour Synapses?
  6. Paper Alert: Microglia Mediate Synaptic Loss in Early Alzheimer’s Disease

Paper Citations

  1. . Apolipoprotein E isoform-dependent amyloid deposition and neuritic degeneration in a mouse model of Alzheimer's disease. Proc Natl Acad Sci U S A. 2000 Mar 14;97(6):2892-7. PubMed.
  2. . Apolipoprotein E promotes astrocyte colocalization and degradation of deposited amyloid-beta peptides. Nat Med. 2004 Jul;10(7):719-26. PubMed.
  3. . Middle-aged human apoE4 targeted-replacement mice show retention deficits on a wide range of spatial memory tasks. Behav Brain Res. 2008 Nov 21;193(2):174-82. PubMed.
  4. . A Dramatic Increase of C1q Protein in the CNS during Normal Aging. J Neurosci. 2013 Aug 14;33(33):13460-74. PubMed.

Further Reading

Primary Papers

  1. . Novel allele-dependent role for APOE in controlling the rate of synapse pruning by astrocytes. Proc Natl Acad Sci U S A. 2016 Sep 6;113(36):10186-91. Epub 2016 Aug 24 PubMed.