Microglia take on various tasks in the brain, including gobbling up amyloid plaques. Now, a study led by Michael Heneka, University of Bonn, Germany, posits that they also help seed and spread plaques. Published in the December 20 Nature, the findings show that large protein complexes released by microglia, known as ASC specks, latch on to Aβ, driving plaque assembly. In an Alzheimer’s mouse model, knocking out ASC, the gene encoding the protein that makes up the speck scaffold, tamped down Aβ pathology and enabled mice to learn and remember better. In both mouse and human brains, the researchers spotted ASC specks in plaque cores. Alzforum covered preliminary data from this study when Heneka presented it last June at Mechanisms of Neurodegeneration, an EMBO/EMBL symposium in Heidelberg, Germany (Jul 2017 conference news). 

  • ASC specks fuel Aβ aggregation.
  • Eliminating ASC protein reduces Aβ burden, rescues memory loss in AD mice.
  • Human Aβ plaque cores harbor ASC proteins.

“[First author Carmen] Venegas and colleagues’ data make a strong case that ASC specks help to promote amyloid-β deposition in APP/PS1 mice,” wrote Richard Ransohoff, Harvard Medical School, Boston, in an accompanying News and Views. To Marco Colonna, Washington University, St. Louis, the work underscores the complexity of microglial biology. “The study emphasizes a multifactorial role for microglia in Alzheimer’s disease; microglia can be good in the context of corralling and containing plaques, but bad if they activate the inflammasome,” Colonna said.

Heneka previously reported that silencing the NLRP3 inflammasome, a protein complex that triggers a cytokine cascade in microglia, protected APP/PS1 mice from learning and memory deficits (Dec 2012 news). The inflammasome orchestrates the activation and release of interleukin-1β and also triggers the release of ASC, short for “apoptosis-associated speck-like protein containing a caspase-recruitment domain.”

At the Core.

ASC specks (red) inhabit the cores of Aβ (green) plaques in the hippocampi of an APP-PS1 mouse (left) and an AD patient (right). [Courtesy of Carmen Venegas.]

Wondering whether the ASC specks were to blame for the inflammasome’s harmful effects, co-first authors Venegas and Sathish Kumar examined how they interacted with Aβ in test tubes, in mice, and in people. They found that ASC specks purified from mouse macrophages accelerated both Aβ40 and Aβ42 aggregation in a time- and dose-dependent manner, shortening the lag phase that characterizes amyloid formation. ASC specks also co-sedimented with synthetic Aβ oligomers and fibrils. 

To pinpoint the ASC domain responsible for the apparent seeding effect, the researchers ran aggregation assays using recombinant ASC variants. Mutations in the PYD domain obliterated ASC’s seeding ability. This region helps ASC oligomerize into helical fibrils that hold the specks together.

In APP/PS1 mice, plaque cores were enriched in ASC protein, even in animals as young as four months (image above). Crossing APP/PS1 mice with ASC knockouts dropped the Aβ load of eight- and 12-month-old offspring to about half that of APP/PS1 mice. The crosses performed better in the Morris water maze.

The researchers detected ASC protein in postmortem human AD plaques in the hippocampus (image above), even in an early stage patient with mild cognitive impairment. Heneka said they did not quantify the fraction of plaques containing ASC, but they hadn’t seen many without. Immunoprecipitation and fractionation experiments using samples from healthy controls and AD patients confirmed the ASC-Aβ association, particularly in plaque cores.

To test whether ASC specks help seed and spread Aβ pathology, the authors turned to injection models. They seeded extracts from plaque-laden mouse brains into one side of the hippocampi of three-month-old APP-PS1 mice. These animals normally develop plaques by around six months. Venegas injected wild-type brain extract into the contralateral hippocampus as a control. As expected from previous findings (Sep 2006 newsMeyer-Leuhmann et al., 2006), after five months, the side of the brain injected with the APP-PS1 extract mustered about twice as many plaques as the contralateral side. Strikingly, the APP-PS1 extract had no effect in APP-PS1 mice lacking ASC. Co-injecting anti-ASC antibodies also blocked seeding by the APP-PS1 extract.

In another experiment, extracts from older APP-PS1;Asc-/- mice evoked only half as many plaques in APP-PS1 mice as did extracts from older APP-PS1 mice.

How important might the contribution of ASC specks be to seeding Aβ plaques in people? Heneka thinks it could be substantial. “ASC specks massively enhance the aggregation propensity of Aβ40, and also of Aβ42, so they have a similar effect to most familial AD mutations,” he said. Risk factors for sporadic AD that activate the inflammasome might initiate Aβ pathology through this ASC-based mechanism, he speculated. Also, the new findings could help explain why synthetic Aβ has proved inefficient as a seed: ASC specks might be the co-factor needed to drive efficient Aβ assembly.

Lary Walker, Emory University, Atlanta, was excited about the idea of ASC specks fueling Aβ aggregation, but suggested a somewhat different scenario. “ASC specks may not be the missing co-factor, but a contributing factor to a complex phenotype. Aβ seeds Aβ very efficiently, and this is likely the backbone of the process, but it also initiates other changes, such as ASC speck release, which could then feed back and accelerate the process,” he said.

Konrad Beyreuther, University of Heidelberg, Germany, and Colin Masters, University of Melbourne, Australia, suggested ASC specks might alter the equilibrium between different forms of Aβ, pushing the balance away from non-aggregating, α-helical, or unfolded forms, and toward aggregation-prone β-structures.

Beyreuther and Masters wondered how to reconcile ASC speck’s enhancement of both Aβ40 and Aβ42 aggregation with Aβ42’s dominant role in AD. “ASC converts Aβ40 into a kind of super-Aβ42,” said Heneka. “In other words, aggregation propensity seen otherwise only with Aβ42, is now being displayed by ASC-bound Aβ40,” he said. Heneka and his collaborators are probing genetic links and so far have found connections between inflammasomes/ASC specks and AD-relevant genes, including APOE. 

Could ASC specks make good drug targets? Although reducing general inflammation has failed as an AD strategy, inflammasomes present new, more specific targets (Sep 2017 news). Walker was optimistic, but emphasized the need for further clarification. “Inflammation is a tough nut to crack—it’s one word but many different processes. If there’s a particular component that could be targeted for AD, that would be great and this study hints at it,” he said.

Indeed, Heneka collaborates with companies, including IFM Therapeutics in Boston, to devise treatments that specifically target ASC specks. Two authors on the paper, Eicke Latz and Matthias Geyer, are co-founders of the company, which develops drugs to target the innate immune system for the treatment of inflammatory disorders and cancer. “The key questions are when [in the course of disease] would be best to intervene, and for how long,” Heneka said.—Marina Chicurel


  1. Based on the identification of full-length Aß42 and N-terminally heterogeneous Aß42 as the main Aß species in amyloid plaques isolated from patients with sporadic Alzheimer’s disease, we have postulated that Aß42 is a key player in AD pathogenesis. Indeed, more recent data confirm that Aß42 species are dominant in the Alzheimer’s disease brain, that it takes 19 years to accumulate amyloid from threshold PET uptake value ratio to the mean value observed for AD dementia, and that the 4.8 mg difference between Alzheimer’s disease and control brain corresponds to an Aß42 accumulation of 28 ng/h (Roberts et al., 2017). The findings from Venegas et al. complement this research by demonstrating that approximately 40–50 percent of these 28 ng/h are controlled by microglia-derived ASC specks. Apoptosis-associated speck-like protein assemblies—ASC specks—undergo a prion-like polymerization upon activation of inflammasome sensors and form large protein complexes reaching a size of around 1 μm, well-suited for simple readout of its role in Aß pathogenesis (Lu et al., 2014).

    As shown by Venegas et al., Aß deposition in AD activates inflammasome-dependent formation of ASC specks in microglia and leads to their subsequent release. In hippocampal sections of AD brains, the majority of ASC specks are detected outside of microglia and the extracellular ASC specks are co-localized with the Aß42 plaques. In APPswePSEN1dE11 mice crossed with ASC-/-, at eight months of age the number of Aß deposits is reduced to half of what is observed for PPswePSEN1dE11 mice. Co-injection of anti-ASC antibodies with APPswePSEN1dE11 brain lysates of eight-month-old mice into three-month APPswePSEN1dE11 mice resulted in approximately 40–50 percent reduction in total number of Aß deposits, again most likely Aß42 aggregates, after five months. Since in these injection experiments relative phagocytosis and levels of the Aß-degrading proteases IDE, NEP, and CASP-1 remained grossly unaltered, degradation of Aß42 is unlikely to account for the observed blocked increase in Aß42 deposition in APPswePSEN1dE11 mice.

    Are ASC specks then the hitherto missing co-factor for efficient induction of Aß42 plaque formation with synthetic Aß? Given the same effect that extracellular ASC speck has in vitro and in vivo on Aß40 and on Aß42 aggregation, and the highly different aggregation properties of both Aß species in vivo (McGowan et al., 2005), it is more likely that the ASC speck interactions influence the equilibrium between non-aggregating, α-helical or unfolded forms of Aß and the aggregation-prone, ß-structure form of Aß. One remaining question is why does Aß40 not form amyloid plaques in the presence of ASC spots? And another question might be whether intracellular ASC-Aß42 complexes form a nidus for other types of aggregates which result in the formation of neurofibrillary tangles and neurites (He et al., 2017). 


    . Biochemically-defined pools of amyloid-β in sporadic Alzheimer's disease: correlation with amyloid PET. Brain. 2017 May 1;140(5):1486-1498. PubMed.

    . Unified polymerization mechanism for the assembly of ASC-dependent inflammasomes. Cell. 2014 Mar 13;156(6):1193-1206. PubMed.

    . Abeta42 is essential for parenchymal and vascular amyloid deposition in mice. Neuron. 2005 Jul 21;47(2):191-199. PubMed.

    . Amyloid-β plaques enhance Alzheimer's brain tau-seeded pathologies by facilitating neuritic plaque tau aggregation. Nat Med. 2018 Jan;24(1):29-38. Epub 2017 Dec 4 PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Microglial Regulation and Function Scrutinized at Heidelberg Meeting
  2. Microglia and AD—Does the Inflammasome Drive Aβ Pathology?
  3. Double Paper Alert—A Function for BACE, a Basis for Amyloid
  4. New AD Target: Silencing the NLRP3 Inflammasome with Boron?

Research Models Citations

  1. APPswe/PSEN1dE9

Paper Citations

  1. . Exogenous induction of cerebral beta-amyloidogenesis is governed by agent and host. Science. 2006 Sep 22;313(5794):1781-4. PubMed.

Further Reading

Primary Papers

  1. . Microglia-derived ASC specks cross-seed amyloid-β in Alzheimer's disease. Nature. 2017 Dec 20;552(7685):355-361. PubMed.
  2. . Specks of insight into Alzheimer's disease. Nature. 2017 Dec 21;552(7685):342-343. PubMed.