The accumulation of phagocytic microglia in the brain could be good or bad for AD patients. Some evidence suggests the microglia can eat up Aβ and clear plaques, but their activation can also lead to harmful neuroinflammation (see ARF related news story). A new report from Joseph El Khoury and colleagues of Harvard Medical School, published this week in the Journal of Neuroscience, finds a way to reconcile these two opposing functions. By tracking gene expression over time in microglia freshly isolated from AD mouse models, El Khoury and coauthors Suzanne Hickman and Elizabeth Allison report that the cells seem to start out by battling the buildup of amyloid. However, over time, the microglia lose their ability to take up and degrade Aβ, and instead show increased expression of inflammatory cytokines. The Alzforum previously reported on some of the data after they were presented last fall in Boston (see ARF related news story).

“We have found in the past that microglia can be both protective and detrimental, and I think this paper shows that at some stage of the disease, the microglia are good and can function to come in and clear amyloid, but either because they get overwhelmed or because of the inflammatory response itself, later they fail,” El Khoury told Alzforum.

The results offer a nuanced view of microglia action in AD, and suggest that anti-inflammatory therapy for the disease should take both sides of their action into account.

The work is based on a technique El Khoury and colleagues developed for rapidly isolating fresh microglia from adult mouse brain using antibodies to the cell surface marker CD11b. To study changes in the microglia as plaque accumulation progresses and microgliosis sets in, the investigators isolated cells from brains of PS1-APP mice at 1.5, three, eight, and 14 months of age, and measured expression of Aβ receptors, Aβ degrading enzymes, and inflammatory markers by quantitative PCR.

By the time they were eight months old, the PS1-APP mice showed significantly reduced expression of the Aβ receptors scavenger receptor A (SRA), CD36, and RAGE compared to their non-transgenic littermates. Wild-type mice also showed reductions in the Aβ receptors, but to a lesser extent, and except for a decrease in CD36, the changes were not statistically significant. In parallel to the decrease in Aβ scavenger receptors, the microglia also showed decreased mRNA for the Aβ-degrading enzymes insulysin, neprilysin, and matrix metalloproteinase 9. In 14-month-old mice, the levels were reduced by 50-80 percent compared to non-transgenic mice. Together, these data could indicate a decreased capacity of the microglia to take up and degrade Aβ, the authors suggest.

In parallel with the apparent decrease in Aβ handling proteins, the microglia from eight- and 14-month-old mice showed increased expression of mRNA for the inflammatory cytokines IL-1β and TNFα. To test the idea that the production of cytokines in a chronic inflammatory response might control expression of Aβ receptors and degrading enzymes, the researchers treated cultured microglia with TNFα and IL-1β. They found that TNFα, but not IL-1β treatment resulted in a significant decrease in SRA and CD36 mRNA and surface expression, as well as Aβ uptake by the cells. On the other hand, neither neprilysin nor insulysin expression was affected by TNFα.

“Our data provide evidence to support the paradigm that the inflammatory response in AD is a ‘double-edged sword,’” the authors write. Initially, newly recruited microglia act to clear Aβ, but with time, they lose their ability to take up and degrade the peptide, while continuing to produce inflammatory cytokines. Not only do the cytokines have the potential to suppress Aβ clearance, but they may also stimulate production of Aβ via their previously documented effects on β-secretase expression and γ-secretase activity.

The reduction in Aβ receptors and degrading enzymes in microglia was initially surprising, El Khoury said. “We were expecting to see an upregulation of genes, because we think that in inflammation you see an increase in a lot of things. But if you look in the literature, people have described in patients as well as in animal models, a reduction in amyloid degrading enzymes where there is more amyloid in the brain, but this was not attributed to microglia. There have also been reports showing that in peripheral macrophages from the blood of Alzheimer’s patients, there may be a reduction in the ability to phagocytose Aβ [see ARF related news story].” These previous results gave hints that the microglia might be losing their ability to combat Aβ buildup, consistent with the new results, El Khoury said.

The results raise the possibility that microglia might be a moving target for anti-inflammatory action. One possibility for the future will be to identify drugs that will “target specific cytokines or receptors without causing an across-the-board downregulation of microglial phagocytosis and degradation of Aβ,” El Khoury said. “Alternatively, therapy could be designed to upregulate microglial clearing ability without inducing inflammatory cytokines.” Work last year from the El Khoury lab (see ARF related news story), and another recent report (see ARF related news story) have both indicated that microglia recruitment to the brain can protect from amyloid buildup in AD mice.

Going forward, El Khoury reports his lab is analyzing expression of many other genes in freshly isolated microglia, cells that were not readily accessible before. They are especially interested in looking at what happens during normal aging, as well as in AD. “There were some studies in the past but they mostly involved culturing the microglia, which doesn’t really give a good snapshot,” El Khoury said. “One advantage of what we are doing is it gives a peek at what we think might be happening in vivo.”—Pat McCaffrey


  1. I’d like to make two major points about this paper from Hickman and colleagues.

    First, I don’t view changing gene expression in microglia as a dysfunction. Microglia change their protein synthesis all the time, including surface receptors and proinflammatory cytokines, especially when activated under conditions of acute injury. The changes reported here for PS1-APP mice seem to be a result of the genetic manipulations in these mice since they do not occur in wild-type cells.

    Second, the role of microglia in amyloid clearance remains controversial. Certainly, the histopathology of human brain does not support a role for microglia in amyloid clearance, especially when it comes to early diffuse amyloid, which does not seem to elicit any kind of response from microglia, i.e., the cells remain in their resting state.

    That said, I do feel that understanding the role of microglia in AD is key to devising new treatments. I also think that microglial dysfunction plays a role, but in the sense that microglia themselves are subject to degeneration, resulting in a dysfunction that manifests itself in a loss of neuroprotection. In other words, I think that neurodegeneration in AD results from neglect (due to a loss of microglia) rather than from aggression (due to neuroinflammation).

  2. The timely report by Hickman, Allison, and El Khoury presents an interesting interpretation of the interplay between microglia and cerebral amyloidosis. It has long been established that Tg2576 mice manifest microglial activation concomitant with Abeta deposition, and that before plaques develop these animals have very little microgliosis (see for example Benzing et al., 1999). These authors have performed a related study in the APPPS1 mice developed by Joanna Jankowsky and David Borchelt (Jankowsky et al., 2001) and find a similar phenomenon.

    They open their abstract by stating that “Early microglial accumulation in Alzheimer’s disease (AD) delays disease progression by promoting clearance of beta-amyloid (Abeta) before formation of senile plaques”. However, I'd like to note that this is a controversial statement, for which the authors do not present experimental evidence. Early ultrastructural studies from Henryk Wisniewski and Jerzy Wegiel actually suggested the opposite, that early microglial activation is a key factor in promoting progression of cerebral amyloidosis (Wisniewski and Wegiel, 1994). Further, observations in Tg2576 mice do not support an association between reactive microglia and diffuse beta-amyloid deposits (only between activated microglia and mature Abeta deposits, Benzing et al., 1999). This seems partially at odds with the authors’ contention.

    Certainly, there are at least two interpretations for the observation of microglial activation occurring in tight temporal and spatial association with more mature beta-amyloid plaques. 1) That, as the authors contend, reactive microglia begin to clear non-deposited (i.e., soluble oligomeric) forms of Abeta and then become easily overwhelmed, or 2) that the microglia become activated in response to “seeds” (e.g., protofibrils or fibrils) of beta-amyloid plaques and then, via chronic, low-level production of pro-inflammatory cytokines and acute-phase reactants, contribute to plaque maturation. Probably the most direct support for the latter interpretation comes from the NSAID epidemiologic literature, where use of NSAIDs is associated with reduced microglial activation in humans (Mackenzie and Munoz, 1998) and as much as 50 percent reduced risk for AD (in t’Veld et al., 2001; Szekely et al., 2004). More recently at ICAD 2008, John Breitner showed impressively that naproxen given during the randomized controlled ADAPT trial for approximately 2 years duration and followed-up for another ~2 years results in reduced incidence of AD.

    But the real “meat” of the work by Hickman et al. comes from their FACS approach as applied to single brain cell suspensions from APPPS1 mice of different ages. Using this approach, the authors show that CD11b+ (presumed microglial) cells from younger APPPS1 animals (from 1.5 to 3 months old, before obvious manifestation of beta-amyloid plaques) appear remarkably similar to age-matched wild-type controls when measuring mRNA for the microglial Abeta uptake receptors SRA, CD36, and RAGE. Similar results where observed on the Abeta-degrading enzymes Insulysin, Neprilysin, and MMP9.

    However, a different pattern of results emerged when considering older (from 8 to 14 months) animals; in this case, APPPS1 mice had significant reductions in both the Abeta phagocytosis receptors and the Abeta degrading enzymes. Interestingly, these same older APPPS1 mice demonstrated up-regulation of mRNA for the pro-inflammatory cytokines IL-1beta and TNF-alpha, suggesting that these microglia are undergoing a phenotype “shift” from Abeta phagocytic, non-inflammatory to Abeta anti-phagocytic, pro-inflammatory.

    We have suggested something similar when we defined microglial activation as a continuum of responses ranging from productive (pro-phagocytic and anti-inflammatory) to deleterious (anti-phagocytic and pro-inflammatory) (Town et al., 2005). We agree with the authors that, to ensure productive clearance of Abeta by phagocytes such as microglia, therapies should promote an anti-inflammatory, pro-phagocytic phenotype. As in-vitro proof-of-concept, the authors show that treatment of N9 microglia with TNF-alpha reduces expression of SRA and CD36 and opposes Abeta uptake by these cells. We have observed something very similar when blocking the pro-inflammatory CD40-CD40L interaction on microglia – we then see increased Abeta uptake and clearance by microglia and reduced pro-inflammatory antigen presenting cell function (Tan, Town et al., 1999; Townsend, Town et al., 2005).

    One further issue that deserves mentioning is the origin of the CD11b+ cells that the authors have nicely characterized. It has now been clearly demonstrated that peripheral macrophages do enter brains of AD mice (Stalder et al., 2005; Simard et al., 2006; El Khoury et al., 2007), and most recently we have shown that boosting their brain entry by blocking TGF-betaRII signaling on these cells reduces AD-like pathology (Town et al., 2008). One wonders what percentage of the CD11b+ cells described by the authors are from the periphery. Nancy Ruddle has routinely used CD45int (brain-resident microglia) versus CD45hi (blood-borne macrophages) FACS staining to discriminate between the two populations (Juedes and Ruddle, 2001), and we have recently employed her protocol for this purpose (Town et al., 2008). The key question remains of whether these blood-borne macrophages are more efficient Abeta phagocytes than their long-term CNS-resident microglial cousins.


    . Evidence for glial-mediated inflammation in aged APP(SW) transgenic mice. Neurobiol Aging. 1999 Nov-Dec;20(6):581-9. PubMed.

    . Ccr2 deficiency impairs microglial accumulation and accelerates progression of Alzheimer-like disease. Nat Med. 2007 Apr;13(4):432-8. PubMed.

    . Nonsteroidal antiinflammatory drugs and the risk of Alzheimer's disease. N Engl J Med. 2001 Nov 22;345(21):1515-21. PubMed.

    . Co-expression of multiple transgenes in mouse CNS: a comparison of strategies. Biomol Eng. 2001 Jun;17(6):157-65. PubMed.

    . Resident and infiltrating central nervous system APCs regulate the emergence and resolution of experimental autoimmune encephalomyelitis. J Immunol. 2001 Apr 15;166(8):5168-75. PubMed.

    . Bone marrow-derived microglia play a critical role in restricting senile plaque formation in Alzheimer's disease. Neuron. 2006 Feb 16;49(4):489-502. PubMed.

    . Invasion of hematopoietic cells into the brain of amyloid precursor protein transgenic mice. J Neurosci. 2005 Nov 30;25(48):11125-32. PubMed.

    . Nonsteroidal anti-inflammatory drugs for the prevention of Alzheimer's disease: a systematic review. Neuroepidemiology. 2004 Jul-Aug;23(4):159-69. PubMed.

    . Microglial activation resulting from CD40-CD40L interaction after beta-amyloid stimulation. Science. 1999 Dec 17;286(5448):2352-5. PubMed.

    . The microglial "activation" continuum: from innate to adaptive responses. J Neuroinflammation. 2005 Oct 31;2:24. PubMed.

    . Blocking TGF-beta-Smad2/3 innate immune signaling mitigates Alzheimer-like pathology. Nat Med. 2008 Jun;14(6):681-7. PubMed.

    . CD40 signaling regulates innate and adaptive activation of microglia in response to amyloid beta-peptide. Eur J Immunol. 2005 Mar;35(3):901-10. PubMed.

    . The role of microglia in amyloid fibril formation. Neuropathol Appl Neurobiol. 1994 Apr;20(2):192-4. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Microglia—Medics or Meddlers in Dementia
  2. Merck Symposium: Surmounting the Blood-brain Barrier in Dementia Research
  3. Clearing Aggregates—Macrophages Fall Short for Aβ, but Vaccine Mops Up α-synuclein
  4. Macrophages Storm Blood-brain Barrier, Clear Plaques—or Do They?

Further Reading

Primary Papers

  1. . Microglial dysfunction and defective beta-amyloid clearance pathways in aging Alzheimer's disease mice. J Neurosci. 2008 Aug 13;28(33):8354-60. PubMed.