Yin C, Ackermann S, Ma Z, Mohanta SK, Zhang C, Li Y, Nietzsche S, Westermann M, Peng L, Hu D, Bontha SV, Srikakulapu P, Beer M, Megens RT, Steffens S, Hildner M, Halder LD, Eckstein HH, Pelisek J, Herms J, Roeber S, Arzberger T, Borodovsky A, Habenicht L, Binder CJ, Weber C, Zipfel PF, Skerka C, Habenicht AJ. ApoE attenuates unresolvable inflammation by complex formation with activated C1q. Nat Med. 2019 Mar;25(3):496-506. Epub 2019 Jan 28 PubMed.

Recommends

Please login to recommend the paper.

Comments

  1. The work identifies a mechanism for the known anti-inflammatory effects of ApoE through a series of approaches. The authors demonstrate that ApoE’s high-affinity binding to the C1q protein inhibits the classical complement cascade, a clear example of a disease-modifying effect of ApoE not directly related to amyloid. The work also shows interesting early pathological changes in the choroid plexus, with an accumulation of lipids and leukocytes at this blood-CSF barrier. Interestingly, these effects are precipitated by a combination of the strong genetic (ApoE4) and environmental (high-fat diet) risk factors for Alzheimer’s disease. The therapeutic usefulness of these findings is demonstrated in the binding of an ApoE-derived peptide to C1q.

    This mechanistic insight fits with previous research indicating ApoE peptides decrease inflammation and provide neuroprotection across several neuropathological conditions. Together the data support a model of loss of ApoE function in the ApoE4 isoform and suggest that research into the detrimental effects of ApoE (in the presence of environmental stresses) is a useful approach to AD-prevention strategies.

    View all comments by G. William Rebeck

  2. This is quite a paper! The authors present convincing evidence that ApoE is a checkpoint inhibitor for complement activation once complement activation is underway. This seems to be a common pathway for both AD and atherosclerosis. This is very interesting, as vascular disease is highly prominent in Alzheimer’s disease. The paper is thorough and intriguing, and really got the wheels turning in my brain.

    Their data with the C5 siRNA treatment in APP PS1 21 mice is quite similar to our data in germline-C3-deficient APP/PS1dE9 mice, in which microglia association with plaques was significantly reduced at 16 months of age. It would be interesting to go back to those mouse brains and look for Aβ/ApoE and C1q/ApoE complexes near or within plaques.

    It would also be interesting to look at brain tissue from people who developed ARIA following Aβ immunotherapy. Antibody-antigen complexes activate the classical complement cascade. Perhaps ApoE4 carriers have diminished ability for ApoE to act as a complement checkpoint inhibitor. This might result in exacerbated chronic inflammation which might then lead to vasogenic edema and/or microhemorrhage following early boluses of antibody, especially in people with the ApoE4 genotype. If so, this would suggest that reducing complement activation—either by modification of antibodies or something like siRNA targeting complement—might be an effective way to combat ARIA side effects in immunotherapy. An interesting question would be whether C1q/ApoE complexes are associated with vascular amyloid in AD ApoE4 versus non-E4 carriers.

    It would be interesting to know whether there are sex differences in the ability of ApoE to inhibit complement. Female ApoE4 carriers seem to be at higher risk of AD, have more Aβ, and progress more quickly than males with the same ApoE4 genotype.  

    Finally, the presence of ApoE/Aβ complexes but lack of ApoE/p-tau complexes in plaques in AD brain seems consistent with a report by Bonham et al. (2016) in which they found a significant interaction between CSF ApoE4 and C3 on amyloid and p-tau, but C3 was only associated with p-tau after accounting for Aβ (in CSF). Notably, there is evidence that C1q can bind directly to Aβ (Webster et al., 1995)!

    References:

    Bonham LW, Desikan RS, Yokoyama JS, Alzheimer’s Disease Neuroimaging Initiative. The relationship between complement factor C3, APOE ε4, amyloid and tau in Alzheimer's disease. Acta Neuropathol Commun. 2016 Jun 29;4(1):65. PubMed.

    Webster S, Glabe C, Rogers J. Multivalent binding of complement protein C1Q to the amyloid beta-peptide (A beta) promotes the nucleation phase of A beta aggregation. Biochem Biophys Res Commun. 1995 Dec 26;217(3):869-75. PubMed.

    View all comments by Cynthia Lemere

  3. The present study highlights the connection between cholesterol, the choroid plexus and immune-brain control communication. Yin et al. attribute a key role to APOE in controlling the choroid plexus (ChP), as a functional interface between the brain and the circulating immune cells. Specifically, the authors demonstrate that ApoE acts as the rheostat regulating the classical complement cascade (CCC) following direct binding to C1q, which according to the authors controls leukocyte infiltration into the brain and inflammation. An important implication is the relevance of lipid homeostasis as a marker for the ChP gateway activity. As an apolipoprotein, the primary role of ApoE is lipid binding and transport. Notably, Yin et al. connected locally impaired lipid homeostasis at the ChP to neuroinflammation. Remarkably, a similar phenotype was observed in mice carrying the AD-associated allele ApoE4 when subjected to a high-fat diet (HFD), but not in mice carrying the ApoE3 allele.

    In light of the new data that were recently generated in connection with the ChP, one of the important findings that emerges from the present study is the observation that ChP shows a Type I interferon signature under conditions that impose APOE dysfunctions. In particular, the balance between type I and type II interferon was found to be crucial for the correct functioning of the ChP, with type I interferon responses characterizing altered physiological states such as aging (Baruch et al., 2014; Deczkowska et al., 2017) and Alzheimer’s disease  (Baruch et al., 2015; Mesquita et al., 2015). In light of this, one might suggest that Type I interferons might be upregulated in AD patients carrying the APOE4 allele, which could explain their more severe disease.

    Stefano Suzzi and Afroditi Tsitsou-Kampeli are co-authors of this comment

    References:

    Baruch K, Deczkowska A, David E, Castellano JM, Miller O, Kertser A, Berkutzki T, Barnett-Itzhaki Z, Bezalel D, Wyss-Coray T, Amit I, Schwartz M. Aging-induced type I interferon response at the choroid plexus negatively affects brain function. Science. 2014 Aug 21; PubMed.

    Deczkowska A, Matcovitch-Natan O, Tsitsou-Kampeli A, Ben-Hamo S, Dvir-Szternfeld R, Spinrad A, Singer O, David E, Winter DR, Smith LK, Kertser A, Baruch K, Rosenzweig N, Terem A, Prinz M, Villeda S, Citri A, Amit I, Schwartz M. Mef2C restrains microglial inflammatory response and is lost in brain ageing in an IFN-I-dependent manner. Nat Commun. 2017 Sep 28;8(1):717. PubMed.

    Baruch K, Rosenzweig N, Kertser A, Deczkowska A, Sharif AM, Spinrad A, Tsitsou-Kampeli A, Sarel A, Cahalon L, Schwartz M. Breaking immune tolerance by targeting Foxp3(+) regulatory T cells mitigates Alzheimer's disease pathology. Nat Commun. 2015 Aug 18;6:7967. PubMed.

    Mesquita SD, Ferreira AC, Gao F, Coppola G, Geschwind DH, Sousa JC, Correia-Neves M, Sousa N, Palha JA, Marques F. The choroid plexus transcriptome reveals changes in type I and II interferon responses in a mouse model of Alzheimer's disease. Brain Behav Immun. 2015 Oct;49:280-92. Epub 2015 Jun 16 PubMed.

    View all comments by Michal Schwartz

  4. Yin and co-workers describe a creative, novel approach to try to understand the physiological role of ApoE and its implications for Alzheimer’s and cardiovascular disease. Using ApoE-knockout and humanized ApoE3- and ApoE4-knock-in mouse models, they examined physiological and immunological changes in the choroid plexus. They report increased expression of interferon pathway genes in the choroid plexus of ApoE knockout mice and that this phenotype is reversed with ApoE3 or 4 reconstitution. This is interesting in that aging seems to induce a similar interferon signature in the choroid plexus of wild-type mice, as shown by Baruch and colleagues, and possibly tying the well-known longevity effects of APOE to this mechanism. Additionally, using binding activity assays, Yin reports that ApoE acts as a checkpoint inhibitor of the classical complement cascade by forming a complex with activated C1q. Excitingly, levels of C1q-ApoE complex seem to correlate with pathology and disease in Alzheimer’s patients and with classical complement cascade activation in atherosclerosis. The authors conclude that the C1q-ApoE complex is a new inhibitor of classical complement activity during disease.

    Overall, this is a very interesting, rigorous, and exciting observation that links ApoE with anti-inflammatory effects and the complement pathway. Activation of the complement system has long been associated with neurodegeneration and Alzheimer’s disease, although this intricate innate immune pathway has likely beneficial and detrimental effects depending on the local environment and the stage of disease. The current study proposes ApoE as an inhibitor of C1q and highlights ApoE as a new regulator for the classical complement pathway in the choroid plexus.

    The study opens many exciting new questions and avenues for more research. How does interferon signaling in the choroid plexus relate to ApoE and brain inflammation and aging in general? How exactly does ApoE regulate complement activation in vivo, and how does this relate to the development of Alzheimer disease? It will also be interesting to identify the cell types involved in this response including the origin of ApoE. Importantly, given attempts to lower ApoE protein levels as a therapy for Alzheimer’s disease, it will be important to monitor activity of the complement cascade and consider these new findings.

    View all comments by Zurine De Miguel

  5. We would like to add comments on a few statements raised in the news story:

    Christine Skerka is quoted as saying she “thinks the blood-CSF barrier was compromised as well, because the researchers detected immunoglobulins in the ChP.” We need to clarify: We did not study the blood-CSF barrier. The detection of immunoglobulins in the ChP does not directly support the statement that the blood-CSF barrier is compromised. There are fundamental differences between the blood-brain barrier (BBB) and the blood-CSF barrier (Ransohoff et al., 2012; Schwartz et al., 2014; Shechter et al., 2013). It is important to note that tight junctions forming the BBB are expressed by endothelial cells in the brain parenchyma, while tight junctions forming the blood-CSF barrier are expressed by the ChP epithelial cells. In sharp contrast, the endothelial cell monolayer of the ChP does not participate in the formation of the blood-CSF barrier because it is fenestrated, and even large molecules including immunoglobulins can pass them to enter the space between the endothelium of the ChP and the epithelial cells. Immunoglobulin deposits in the space between the fenestrated endothelial cell monolayer and the epithelial cell monolayer of the ChP are likely unrelated to the blood-CSF barrier. Of course, it is conceivable that the blood-CSF barrier may nevertheless be compromised, but our work does not address or support the speculation by Skerka. Following the lead of Bell et al., however, we observed that the BBB in ApoE-/- and of ApoE4 knock-in mice is compromised (Extended Figure 1e of our manuscript; Bell et al., 2012) and that the BBB breakdown in the brain parenchyma is a function of either the absence of ApoE or the presence of ApoE4. Further studies on the blood-CSF barrier are needed to clarify these important issues.

    Regarding the question of how ApoE isoforms contribute to AD risk based on the data presented in the manuscript, the story quotes Yadong Huang as saying, “That’s the key unresolved question here.” We reported two ApoE4 isoform-specific pro-inflammatory activities in our manuscript (see below points ii, iii).  ApoE4-KI mice on normal chow are normolipidemic and do not have lipid deposits in the ChPs, hence, this data does not directly support the speculation that “the isoforms modulate AD risk by differentially influencing complement activators, such as lipids and Aβ,” as Skerka is quoted as saying. Instead, we propose a different scenario:

    1. Our data support the view that ApoE4 regulates inflammation via an inhibitory action on activated C1q by downregulating the classical complement cascade activity in vitro and in vivo.
    2. We find a major component of ApoE isoform specificity on ChP lipid deposits: HFD-fed ApoE4-KI mice showed a much higher extent of lipid in ChPs when compared to HFD-fed ApoE3-KI mice. We agree with Bill Rebeck that ApoE4, together with hyperlipidemia, causes ChP lipid deposits.
    3. There is another major ApoE4 isotype-specific, pro-inflammatory activity that is unrelated to hyperlipidemia, lipid deposits, and complement pathways or the C1q-ApoE4 complex: Both normal chow (NC) and HFD-fed ApoE4-KI mice showed marked interferon (IFN) signatures in the ChP when compared to ApoE3-KI or WT mice. Clearly, the majority of IFN-related genes appear to be the result of ApoE4 rather than the result of hyperlipidemia (Fig.1 of the current manuscript). We thank Michal Schwartz and Tony Wyss-Coray/De Miguel for emphasizing the large ChP IFN signature. As they point out, it will be of major interest to identify the relation of the ApoE4-related IFN signatures in the ChP and AD pathogenesis in future studies.   

    Based on the above arguments, we propose that the Janus-headed nature of ApoE4 reveal both anti-inflammatory activities by forming the C1q-ApoE complex (hereby inhibiting the classical complement cascade) and pro-inflammatory activities by triggering a strong IFN signature.

    We agree with Huang that the relationship between ApoE and inflammation, and the role of inflammation in neurodegeneration, is not clear-cut. Our study does not clarify this issue, and much work needs to be done to better understand this question. These studies need to take into account major parameters of AD including age, the seeding events during early AD stages, inflammation and its role during AD plaque formation, and synaptic pruning (which may be different during distinct disease stages), and many more, including the isoforms of ApoE. Indeed, recent studies by several groups have indicated that ApoE has opposite effects during the different stages of AD development in mice (Huynh et al., 2017; Liu et al., 2017). 

    We live in exciting times and hopefully are close to uncovering some AD mysteries, but we are not yet there by any account. Our study adds a little mosaic stone to a big picture that is only beginning to be faintly seen in the distance.

    References:

    Bell RD, Winkler EA, Singh I, Sagare AP, Deane R, Wu Z, Holtzman DM, Betsholtz C, Armulik A, Sallstrom J, Berk BC, Zlokovic BV. Apolipoprotein E controls cerebrovascular integrity via cyclophilin A. Nature. 2012 May 24;485(7399):512-6. PubMed. Correction.

    Huynh TV, Liao F, Francis CM, Robinson GO, Serrano JR, Jiang H, Roh J, Finn MB, Sullivan PM, Esparza TJ, Stewart FR, Mahan TE, Ulrich JD, Cole T, Holtzman DM. Age-Dependent Effects of apoE Reduction Using Antisense Oligonucleotides in a Model of β-amyloidosis. Neuron. 2017 Dec 6;96(5):1013-1023.e4. PubMed.

    Liu CC, Zhao N, Fu Y, Wang N, Linares C, Tsai CW, Bu G. ApoE4 Accelerates Early Seeding of Amyloid Pathology. Neuron. 2017 Dec 6;96(5):1024-1032.e3. PubMed.

    Ransohoff RM, Engelhardt B. The anatomical and cellular basis of immune surveillance in the central nervous system. Nat Rev Immunol. 2012 Sep;12(9):623-35. Epub 2012 Aug 20 PubMed.

    Schwartz M, Baruch K. The resolution of neuroinflammation in neurodegeneration: leukocyte recruitment via the choroid plexus. EMBO J. 2014 Jan 7;33(1):7-22. Epub 2013 Dec 19 PubMed.

    Shechter R, London A, Schwartz M. Orchestrated leukocyte recruitment to immune-privileged sites: absolute barriers versus educational gates. Nat Rev Immunol. 2013 Mar;13(3):206-18. PubMed.

    View all comments by Changjun Yin

  6. This is a very interesting series of experiments, conclusively demonstrating an interaction between C1q and ApoE, with only slight differences in affinity for C1q among the ApoE isoforms. The authors then provide evidence of an ApoE peptide that prevents this interaction and, since it does not include the regions differing between the ApoE isoforms, this supports the demonstrated lack of difference in affinity between C1q and ApoE isoforms. [However, use of a scrambled peptide or reverse sequence for the 139-152 peptide would have provided more definitive conclusions.] These data add to the growing body of observations providing a link between ApoE and C1q, such as early work by Eikelenboom (Zhan et al., 1995) and later Gareth Howell’s group (Soto et al., 2015). The data also show the ability of β-amyloid fibrils and not soluble amyloid to activate complement, which reproduces and supports previously published work from our lab (Jiang et al., 1994) and others (Tacnet-Delorme et al., 2001). 

    In addition and importantly, the data presented demonstrate that a knockdown of C5 synthesis using novel siRNA reduced plaque-associated microglia in a mouse model of AD. These data are consistent with a critical detrimental role of C5a in the development of cognitive loss in mouse models of AD as we published in 2009 and 2017 (Hernandez et al., 2017; Fonseca et al., 2009), and thus supports the potential for therapeutic targeting of C5a or its receptor C5aR1.

    However, to avoid confusion and therefore facilitate progress toward effective therapeutic development, it would be helpful to accurately describe the complement components C1q and C1. There is no “inactive” form of C1q, as mentioned in the paper and the Alzforum report above. In the absence of C1r and C1s, or after the dissociation of the activated enzymes C1r and C1s from C1q (which occurs rapidly in blood due to C1 inhibitor), C1q has many activities as described in the review (Thielens et al., 2017). When C1q is complexed with two molecules each of C1r and C1s in C1, C1 is “unactivated”, but "activatable" when referring to its ability to initiate the CCC. (For some readers, inactive can mean misfolded, inhibited, etc.)

    In blood/serum, most of the C1q (90 percent) is found in complex with the proenzymes C1r and C1s (Ziccardi and Tschopp, 1982)—this is “unactivated” C1. When this C1 binds to “activators,” i.e., those molecules that constrain C1q to enable C1r and C1s to be permanently cleaved to active enzymes, the activated C1s (of the now “activated C1”) propagates the complement pathway activity by cleaving the next proteins in the cascade as illustrated in Extended Data Fig. 3 in this manuscript (except that C4 is cleaved first, then C2). [In contrast to what is stated in this manuscript, while C1 is slightly more stable when bound to an activator, C1q bound to a target does not recruit C1r and C1s to any greater extent than free C1q.]

    Furthermore, Ca++, which is in all biological fluids, including NHS, is required for binding of C1r2C1s2 to C1q, but not for the “conformational change” induced by the activator to which C1q binds. Therefore, the Ca++ dependence of ApoE binding to C1q does not indicate “that ApoE selectively binds to the activated form, but not the inactivated form of C1q” as concluded in this paper, but rather that ApoE may bind to “free” C1q in a Ca++ dependent manner. Further explorations of the binding of ApoE to free C1q versus C1q in complex with C1r and C1s (C1) are warranted to determine the nature of these interactions in serum and tissue environments.

    Deborah A. Fraser, California State University, Long Beach, is a co-author of this comment.

    References:

    Zhan SS, Veerhuis R, Kamphorst W, Eikelenboom P. Distribution of beta amyloid associated proteins in plaques in Alzheimer's disease and in the non-demented elderly. Neurodegeneration. 1995 Sep;4(3):291-7. PubMed.

    Soto I, Graham LC, Richter HJ, Simeone SN, Radell JE, Grabowska W, Funkhouser WK, Howell MC, Howell GR. APOE Stabilization by Exercise Prevents Aging Neurovascular Dysfunction and Complement Induction. PLoS Biol. 2015 Oct;13(10):e1002279. Epub 2015 Oct 29 PubMed.

    Jiang H, Burdick D, Glabe CG, Cotman CW, Tenner AJ. beta-Amyloid activates complement by binding to a specific region of the collagen-like domain of the C1q A chain. J Immunol. 1994 May 15;152(10):5050-9. PubMed.

    Tacnet-Delorme P, Chevallier S, Arlaud GJ. Beta-amyloid fibrils activate the C1 complex of complement under physiological conditions: evidence for a binding site for A beta on the C1q globular regions. J Immunol. 2001 Dec 1;167(11):6374-81. PubMed.

    Hernandez MX, Jiang S, Cole TA, Chu SH, Fonseca MI, Fang MJ, Hohsfield LA, Torres MD, Green KN, Wetsel RA, Mortazavi A, Tenner AJ. Prevention of C5aR1 signaling delays microglial inflammatory polarization, favors clearance pathways and suppresses cognitive loss. Mol Neurodegener. 2017 Sep 18;12(1):66. PubMed.

    Fonseca MI, Ager RR, Chu SH, Yazan O, Sanderson SD, LaFerla FM, Taylor SM, Woodruff TM, Tenner AJ. Treatment with a C5aR antagonist decreases pathology and enhances behavioral performance in murine models of Alzheimer's disease. J Immunol. 2009 Jul 15;183(2):1375-83. Epub 2009 Jun 26 PubMed.

    Thielens NM, Tedesco F, Bohlson SS, Gaboriaud C, Tenner AJ. C1q: A fresh look upon an old molecule. Mol Immunol. 2017 Sep;89:73-83. Epub 2017 Jun 7 PubMed.

    Ziccardi RJ, Tschopp J. The dissociation properties of native C1. Biochem Biophys Res Commun. 1982 Jul 30;107(2):618-23. PubMed.

    View all comments by Andrea Tenner

  7. I notice that everyone is largely in agreement with the findings of the authors. All very exciting. I wish to add that the very sites linked with ApoE and complement activation (atherosclerosis, choroid plexus, amyloid-β plaques) are also known to contain pathogenic bacteria and bacterial toxins. Nobody has mentioned their expert opinion on this in relation to the ApoE-C1q checkpoint story. I have worked on ApoE knockout mouse brains and did not find rampant complement activation. These mice were fed on standard chow. In fact, I only found activated complement after confirming entry of bacteria into the brains of ApoE KO mice. Does this mean that complement activation becomes deregulated in ApoE KO mice fed a high fat diet?

    View all comments by Sim Singhrao

Make a Comment

To make a comment you must login or register.

This paper appears in the following:

News

  1. ApoE Binds Complement Protein, Keeps Inflammatory Cascade in Check