Variants in the microglial receptor TREM2 are among the strongest risk factors for late-onset Alzheimer’s disease, but how they contribute to disease had been unclear. Now, some consensus is emerging that TREM2 helps protect the brain. Two new papers in the March 7 Neuron strengthen this idea and clarify what the receptor does. Researchers led by X. William Yang at the David Geffen School of Medicine at the University of California, Los Angeles, overexpressed human TREM2 in an AD mouse model and found that the gene reprogrammed microglia into a more phagocytic and less inflammatory state. The mice developed less amyloid plaque and dystrophic neurites than AD controls, and maintained normal learning and memory. A separate study that involved some of the same authors provides more clues to how TREM2 does this. Researchers led by Huaxi Xu at the Sanford Burnham Prebys Medical Discovery Institute in La Jolla, California, reported that TREM2 directly binds Aβ. This binding triggers microglia to activate, secrete cytokines, and degrade the internalized peptide.

  • AD mice overexpressing TREM2 have less amyloid and better memory.
  • Aβ directly binds TREM2 and activates microglia to degrade the peptide.
  • The data suggest that boosting TREM2 could ameliorate AD pathology.

Other researchers said these are important papers that advance the field. “This is the first hint that it is indeed possible to prevent or delay the onset of AD or to modulate the severity of the disease by augmenting the levels of TREM2,” Rita Guerreiro at University College London wrote to Alzforum (see comment below). In an accompanying editorial, Joe Udeochu, Faten Sayed, and Li Gan at the Gladstone Institute of Neurological Disease in San Francisco concurred. “Taken together, these findings support development of TREM2 enhancers as therapeutic strategies to protect against amyloid-associated toxicity in AD,” they wrote.

Previous studies have examined TREM2’s role by characterizing mice that lack the protein. Without TREM2, microglia remain in a quiescent state, migrate and proliferate slowly, and die easily (May 2017 news; Aug 2017 news). Few studies have investigated the effects of TREM2 overexpression in mouse models, though Chinese researchers recently reported that lentiviral expression of mouse TREM2 ameliorated pathology in young mouse models of AD and Parkinson’s disease, but not in aged AD mice (Jiang et al., 2014; Jiang et al., 2017; Ren et al., 2018). 

TREM2 Remodels Microglia.

In AD mice (top), microglia surrounding plaques (red) become reactive, expressing high Iba1 (green) and assuming an amoeboid shape. With higher TREM2 levels (bottom), microglia have longer processes and express less Iba1. [Courtesy of Neuron, Lee et al.]

Yang and colleagues took a different approach to overexpression. Joint first authors C. Y. Daniel Lee and Anthony Daggett generated a mouse model that expressed human TREM2 on a bacterial artificial chromosome (BAC). This allowed the researchers to include 169 kilobases of surrounding genetic material, including the regulatory elements that restrict TREM2 expression to myeloid cells. Importantly, the authors knocked out other TREM-like genes in this region to avoid confounding effects. Using a GFP reporter, the authors confirmed that mice carrying the BAC expressed human TREM2 only in microglia in the brain. In young, healthy mice, just 6 to 9 percent of microglia expressed the protein.

The picture changed when the BAC-TREM2 mice were crossed with 5xFAD mice, which model aggressive amyloidosis. In 5xFAD mice, endogenous TREM2 levels rise around seven months of age, five months after the first plaques appear. In the BAC-TREM2/5xFAD crosses, human TREM2 expression rose earlier, at four months, and higher at seven. “The model recapitulates the disease-associated upregulation of TREM2,” Yang noted.

The TREM2 overexpressers fared better than 5xFAD mice. Their plaque load was lower by about one-fourth, and plaques appeared more compact and less filamentous. Other studies suggest that TREM2 helps microglia wall off and contain plaques (May 2016 news). Meanwhile, the crossed mice had only about a third as many dystrophic neurites as the 5xFAD mice, and their performance on a fear conditioning test resembled that of wild-types. The authors saw similar results in a second mouse model. APPswe/PSEN1dE9 mice crossed with BAC-TREM2 animals remembered an aversive stimulus nearly as well as wild-types.

Notably, BAC-TREM2/5xFAD mice had fewer microglia around plaques than the 5xFAD animals did. This puzzled many commentators, because TREM2 is believed to encourage microglia to migrate toward plaques. In TREM2 knockouts, very few microglia surround plaques. What might be going on in the overexpressers? Yang noted that plaque-associated microglia in the crossed animals looked different. Instead of having an amoeboid shape with short processes, characteristic of reactive microglia, they maintained long, branched processes. In addition, they expressed less of the reactive marker Iba1 (see image above). Plaque-associated microglia in the APP/PS1 TREM2 overexpressers likewise took on this elongated, branched appearance. Yang wondered if it represented a different activation state.

To glean some clues, the authors analyzed gene expression in whole cortical samples from TREM2/5xFAD and 5xFAD mice at two, four, and seven months of age. About 50 genes were differentially expressed in the two models. These fell into three categories. The first were inflammatory microglial genes that surged with age in 5xFAD mice but were partially suppressed in the TREM2 overexpressers. This group included several AD-linked genes, such as CH25h, ABI3, and TREM2 itself. A second group consisted of synaptic and neuronal genes that were low in 5xFAD mice but closer to wild-type levels in the TREM2/5xFADs, suggesting a normalization of neuronal function. The third group were expressed by 5xFAD mice at nearly wild-type levels, but spiked in the TREM2 overexpressers. These included genes involved in phagocytosis and inhibition of T cell activation, suggesting the microglia enter clean-up mode.

The authors tested this idea by staining for phagocytic markers in plaque-associated microglia. In TREM2/5xFAD mice, microglia expressed more CD68, a marker for lysosomal digestion, than they did in 5xFAD animals. They also pumped out a surfeit of Lgals3, a secreted lectin that binds debris and acts as an “eat me” signal. To further investigate phagocytic capability, the authors isolated microglia from newborn wild-type, and TREM2 knockout mice. The TREM2 knockout microglia poorly phagocytosed beads in culture, while BAC-TREM2 mice ate voraciously, in agreement with previous studies suggesting that TREM2 stimulates phagocytosis (Jul 2016 news). 

“Our interpretation is that TREM2 reprograms microglia, making them more efficient at removing or walling off plaques,” Yang told Alzforum. “This is the first solid in vivo evidence that by fine-tuning the TREM2 level, and doing so early enough, you can change the way microglia respond to disease.” In future work, he plans to further explore the role of downstream genes in this reprogramming. He cautioned, however, that these mouse models did not express abnormal tau, thus it was unclear what the net effect of more TREM2 would be in the AD brain. Some studies have found that TREM2 can exacerbate tau pathology, while others report improvement (Oct 2017 news). Yang will explore the effects of TREM2 overexpression in tauopathy models as well as in models of other neurodegenerative diseases with an inflammatory component, such as Huntington’s.

Other researchers applauded the findings. “The paper by Lee et al. is a momentous step forward for understanding how TREM2 function governs AD-related pathology,” David Hansen at Genentech, South San Francisco, wrote to Alzforum (see comment below). Monica Carson at the University of California, Riverside, agreed that the overexpression model helps answer questions. “This was the critical experiment that had to be done,” she told Alzforum. She noted that although TREM2 rises along with amyloid pathology in model mice and AD brains, these data suggest that boosting TREM2 earlier would help the brain fight disease. Gernot Kleinberger at the German Center for Neurodegenerative Diseases (DZNE) in Munich concurred. “This study further argues in favor of a beneficial rather than detrimental role of TREM2, especially in the early phases of amyloid plaque deposition,” he wrote (see comment below).

At the same time, researchers noted some caveats. Oleg Butovsky at Brigham and Women’s Hospital, Boston, pointed out that analyzing gene expression from cortices could introduce confounds, especially because the relative proportion of microglia can change in the AD brain. Marco Colonna and Wilbur Song at Washington University in St. Louis, Missouri, noted that the mouse model expressed both human and mouse TREM2, which may have competed with each other to bind the DAP12 co-receptor. Colonna suggested repeating the overexpression study in a TREM2 knockout mouse to avoid potential confounding effects.

The TREM2 overexpression study leaves open the question of what causes the receptor to react to amyloid. The study by Xu and colleagues suggests that Aβ itself triggers this. Joint first authors Yingjun Zhao, Xilin Wu, Xiaoguang Li, and Lu-Lin Jiang prepared synthetic Aβ42 monomers and oligomers, and found that the extracellular portion of TREM2 bound to the oligomers, but not monomers, with a similar affinity to that of known Aβ receptors. The authors confirmed the interaction using bio-layer interferometry assays and surface plasmon resonance assays. They also immunoprecipitated an Aβ/TREM2 complex from brain lysates taken from an AD model as well as AD tissue samples, suggesting the interaction occurs in vivo as well.

Notably, the AD-associated TREM2 variants R47H and R62H bound poorly to oligomeric Aβ. Commenters said the finding needs to be replicated, but could be key. “It provides a basis for the R47H variant affecting AD but not other neurodegenerative diseases,” Colonna said. Butovsky agreed, “This is really exciting if true.”

This paper is the first published report of Aβ binding to TREM2, but it matches unpublished work discussed by Peter St. George-Hyslop at the 2017 AD/PD conference (Apr 2017 conference news). St. George-Hyslop also reported weaker binding of Aβ oligomers by the R47H variant. Likewise, another study found that soluble R47H TREM2 poorly binds amyloid plaques (Jan 2018 news). However, a forthcoming paper from researchers led by Todd Golde at the University of Florida, Gainesville, currently in preprint form on BioRxiv, confirms that TREM2 recognizes Aβ oligomers, but reports that the R47H variant binds them with equal affinity. Instead of impaired binding, Golde and colleagues found weak NFAT signaling through R47H after Aβ binding.

“This will be an important discrepancy to solve,” noted Colonna, who is a co-author on Golde’s paper.

What are the consequences of Aβ binding TREM2? Zhao and colleagues compared microglia isolated from newborn wild-type and TREM2 knockout mice. They found that the TREM2 knockouts took up Aβ as well as the wild-types, but were unable to degrade it. In the wild-type microglia, the authors were able to prevent Aβ degradation with proteasome but not lysosome inhibitors, suggesting the cells eliminated most of the peptide through the former process. TREM2 knockout microglia also failed to activate and secrete pro-inflammatory cytokines in response to Aβ.

To see if this held true in vivo, the authors injected Aβ into the brains of wild-type and TREM2 knockout mice. In the knockouts, fewer microglia clustered at the injection site, and they cleared less of the peptide. Microglia also failed to proliferate in response to Aβ.

“TREM2 may function as an Aβ sensor in microglia,” Zhao suggested. He believes that turning up Aβ clearance through the TREM2 pathway may represent an alternative therapeutic strategy to lowering production with BACE inhibitors. He is currently screening small molecules for those that bind and activate TREM2 and could be potential therapeutic candidates.

“Strikingly, both studies provide evidence that TREM2 alters the degradative process in microglia, albeit through disparate mechanisms,” Udeochu, Sayed, and Gan noted in their editorial.—Madolyn Bowman Rogers


  1. These are very significant studies. The results by Lee and colleagues show that extra TREM2 gene dosage is beneficial for memory deficits and AD-associated neuropathology in AD mouse models. This is the first step for us to be able, perhaps, to use increased levels of TREM2 therapeutically. A lot needs to be done, but this is the first hint that it may be possible to prevent or delay the onset of AD, or to modulate the severity of the disease, by augmenting the levels of TREM2. When we identify a new gene contributing to a disease this is the ultimate goal, and it’s great to see the amazing work that has been done by many groups trying to understand how TREM2 is functionally acting in AD and other neurodegenerative diseases.

    The authors have also performed transcriptomic profiling to identify genes dependent on TREM2 gene dosage. This work helps us geneticists to focus on specific genes, so it works both ways.

    The study by Zhao and colleagues reminds me of a cartoon diagram we put together after confirming TREM2 p.R47H as a risk factor for AD, where we speculated what we thought was happening at the cellular level in AD patients harboring TREM2 variants. The finding that TREM2 directly binds to Aβ oligomers may be the missing initial step in that process. One aspect we included in the diagram was the possibility of other risk genes also having effects in microglia and inflammatory activation. Now the authors raise the possibility of TREM2 mediating Aβ catabolism through proteasomal degradation pathways, which could bring together other risk genes known to have functions at this level.

  2. These are really exciting days, with increasing numbers of very interesting and informative studies on the function of TREM2 in the context of Alzheimer’s disease. Following an earlier study by the Colonna lab (Song et al., 2018), Lee and colleagues developed a human TREM2 overexpression mouse model using a BAC-transgenic approach. Importantly, the authors designed their mouse model in a way to only overexpress TREM2, excluding confounding effects by inactivating the adjacent TREM-like genes.

    It has to be highlighted that this study further argues in favor of a beneficial rather than detrimental role of TREM2, especially in the early phases of amyloid plaque deposition. Additionally, the study provides a large set of data to further strengthen the role of TREM2 in modulating the phagocytic capacity of microglia, as has been shown by many groups, including ours (Takahashi et al., 2005; Hsieh et al., 2009; Kleinberger et al., 2014; Xiang et al., 2016). 

    These data might also indicate that the increase of soluble TREM2 in human patients approximately five years before symptom onset (Suárez-Calvet et al., 2016; Suárez-Calvet et al., 2016) represents a beneficial response of microglia to clear amyloid plaques. Based on the consistent data that TREM2 loss of function reduced clustering of microglia around plaques and the strong upregulation of TREM2 in plaque-associated microglia (Frank et al., 2008; Jay et al., 2015) one was tempted to speculate that overexpression of TREM2 would increase plaque-associated microglia. Surprisingly, this is not the case, and even a reduced number of Iba1-positive microglia are clustered around plaques. While the number of Iba1+ microglia is reduced, the number of microglia with increased expression of phagocytic markers (e.g. CD68) is increased around amyloid plaques, calling for attention to include additional markers besides Iba1 in evaluating plaque-associated microglia in future studies.

    Lee and colleagues further nicely show that overexpression of human TREM2 reduces neuritic dystrophy and finally also improves cognitive outcome, as measured by the contextual fear-conditioning test. As it is known that TREM2 is processed by regulated intramembrane proteolysis, releasing its ectodomain into the extracellular space (Kleinberger et al., 2014; Wunderlich et al., 2013), it is more than tempting to speculate that sTREM2 has a non-cell autonomous function. This will remain a major question to be answered in future studies.

    First evidence for a non-cell autonomous function of sTREM2 was presented earlier by Song et al., who similarly used a BAC-transgenic approach demonstrating sTREM2 staining in neurons by immunohistochemistry. Here the authors also observe hTREM2 staining on a fraction of Aβ plaques, leaving open the question whether Aβ is yet another ligand for TREM2.

    The study by Zhao and colleagues kicks in using a panel of in-vitro assays to demonstrate binding of oligomerized Aβ to human TREM2. Especially the data showing that Trem2 regulates electrophysiological changes in microglia upon stimulation with oligomeric Aβ are very interesting. Whether these effects are directly caused by the absence of Trem2 and hence a reduced binding of oligomeric Aβ, or due to a general locking of Trem2-deficient microglia in their homeostatic state (Mazaheri et al., 2017) requires further study.

    Overall, the presented studies point toward a beneficial role of TREM2 in early stages of amyloid deposition and strongly argue for increasing efforts to find ways to modulate TREM2 as a potential therapeutic strategy in neurodegenerative diseases like Alzheimer’s.


    . Humanized TREM2 mice reveal microglia-intrinsic and -extrinsic effects of R47H polymorphism. J Exp Med. 2018 Mar 5;215(3):745-760. Epub 2018 Jan 10 PubMed.

    . Clearance of apoptotic neurons without inflammation by microglial triggering receptor expressed on myeloid cells-2. J Exp Med. 2005 Feb 21;201(4):647-57. PubMed.

    . A role for TREM2 ligands in the phagocytosis of apoptotic neuronal cells by microglia. J Neurochem. 2009 May;109(4):1144-56. Epub 2009 Mar 19 PubMed.

    . TREM2 mutations implicated in neurodegeneration impair cell surface transport and phagocytosis. Sci Transl Med. 2014 Jul 2;6(243):243ra86. PubMed.

    . TREM2 deficiency reduces the efficacy of immunotherapeutic amyloid clearance. EMBO Mol Med. 2016 Sep 1;8(9):992-1004. PubMed.

    . sTREM2 cerebrospinal fluid levels are a potential biomarker for microglia activity in early-stage Alzheimer's disease and associate with neuronal injury markers. EMBO Mol Med. 2016 May 2;8(5):466-76. PubMed.

    . Early changes in CSF sTREM2 in dominantly inherited Alzheimer's disease occur after amyloid deposition and neuronal injury. Sci Transl Med. 2016 Dec 14;8(369):369ra178. PubMed.

    . TREM2 is upregulated in amyloid plaque-associated microglia in aged APP23 transgenic mice. Glia. 2008 Oct;56(13):1438-47. PubMed.

    . TREM2 deficiency eliminates TREM2+ inflammatory macrophages and ameliorates pathology in Alzheimer's disease mouse models. J Exp Med. 2015 Mar 9;212(3):287-95. Epub 2015 Mar 2 PubMed.

    . Sequential proteolytic processing of the triggering receptor expressed on myeloid cells-2 (TREM2) by ectodomain shedding and γ-secretase dependent intramembranous cleavage. J Biol Chem. 2013 Nov 15;288(46):33027-36. PubMed.

    . TREM2 deficiency impairs chemotaxis and microglial responses to neuronal injury. EMBO Rep. 2017 Jul;18(7):1186-1198. Epub 2017 May 8 PubMed.

  3. It is encouraging that elevating TREM2 gene expression resulted in favorable reductions in neuritic dystrophy. This suggests that there is sufficient bandwidth in microglia to augment TREM2 function as a means of modulating the immune response to amyloid pathology. The suggestion that Trem2 overexpression modifies the gene expression profile of plaque-associated microglia is intriguing, and may portend unexpected phenotypes if TREM2 is overexpressed in models of tauopathy. Given the robust expression of microglial ApoE observed in the Krasemann and Keren-Shaul studies, I would be curious to know if TREM2 overexpression affected ApoE in 5xFAD mice.

  4. The paper by Lee et al. is a momentous step forward for understanding how TREM2 function governs AD-related pathology. Genetic approaches in humans and mice had collectively shown that partial or complete loss of TREM2 function enhanced Alzheimer's risk, but evidence for a protective effect of augmented TREM2 function was yet minimal.

    Carrasquillo et al. (2017) reported that an SNP associated with higher expression levels of TREM2 and TREML1 also associated with reduced AD risk, which was suggestive, but the Lee et al. paper is the first to demonstrate a causal link between TREM2 gain of function and mitigation of amyloid-driven AD pathology in vivo.

    An interesting question is whether the protective effects exerted by human TREM2 expression were simply a quantitative effect of increased gene dosage, or whether there were qualitative differences in the ability of human versus murine TREM2 to elicit a more beneficial microglial response. Indeed, certain microglial phenotypes known to depend on murine TREM2 expression, such as congregation around plaque or decreases in microglial process length and branch number, were partially allayed, not further heightened, by human TREM2 expression. However, other interpretations involving dynamic interplay between components of the microglial response that are directly vs. indirectly TREM2-dependent could also explain these results.

    Zhao et al. proposed the unique and interesting finding that TREM2 and Aβ directly interact. If we assume that no impurities/co-purifying factors in the TREM2-Fc protein preps were involved, then the direct interaction deserves further experimental support since Aβ42 is a notoriously sticky peptide with 26 aliphatic and/or bulky hydrophobic residues. In our experience, the TREM2 extracellular domain is also sticky, with several surface-exposed hydrophobic residues.

    Given these concerns, the paper would have been stronger if the authors supported the idea of direct TREM2/Aβ interaction in their cell-based and in vivo experiments. In these settings, the TREM2-dependent effects of Aβ treatment could have occurred through indirect interactions—notably, through lipoproteins which may convey Aβ to microglia and enhance TREM2-dependent uptake, as reported by colleagues here at Genentech (Yeh et al., 2016). 

    It would be nice to know whether the reported effects of Aβ in primary microglial and BV-2 cultures are observed in the absence of lipoproteins, perhaps by using serum-free media and ApoE knockout cells. This question is pertinent given the recent reports of Krasemann/Butovsky and Ulrich/Holtzman, indicating that the normal, TREM2-dependent microglial response to Aβ requires ApoE. If the direct interaction of Aβ with TREM2 is truly key for the microglial response to injected Aβ, the response should still occur to some degree in ApoE knockout mice, so hopefully this idea will be tested in the future.


    . A candidate regulatory variant at the TREM gene cluster associates with decreased Alzheimer's disease risk and increased TREML1 and TREM2 brain gene expression. Alzheimers Dement. 2017 Jun;13(6):663-673. Epub 2016 Dec 8 PubMed.

    . TREM2 Binds to Apolipoproteins, Including APOE and CLU/APOJ, and Thereby Facilitates Uptake of Amyloid-Beta by Microglia. Neuron. 2016 Jul 20;91(2):328-40. PubMed.

Make a Comment

To make a comment you must login or register.


News Citations

  1. Paper Alert: TREM2 Crucial for Microglial Activation
  2. Without TREM2, Microglia Run Out of Gas
  3. Barrier Function: TREM2 Helps Microglia to Compact Amyloid Plaques
  4. TREM2 Helps Phagocytes Gobble Up Aβ Coated in Antibodies
  5. Changing With the Times: Disease Stage Alters TREM2 Effect on Tau
  6. New Evidence Confirms TREM2 Binds Aβ, Drives Protective Response
  7. New Mouse Models Reveal Unexpected Property of TREM2

Research Models Citations

  1. 5xFAD
  2. APPswe/PSEN1dE9
  3. TgCRND8

Paper Citations

  1. . Upregulation of TREM2 ameliorates neuropathology and rescues spatial cognitive impairment in a transgenic mouse model of Alzheimer's disease. Neuropsychopharmacology. 2014 Dec;39(13):2949-62. Epub 2014 Jul 22 PubMed.
  2. . TREM2 Overexpression has No Improvement on Neuropathology and Cognitive Impairment in Aging APPswe/PS1dE9 Mice. Mol Neurobiol. 2016 Jan 16; PubMed.
  3. . TREM2 overexpression attenuates neuroinflammation and protects dopaminergic neurons in experimental models of Parkinson's disease. Exp Neurol. 2018 Jan 27;302:205-213. PubMed.

External Citations

  1. BioRxiv

Further Reading

Primary Papers

  1. . Elevated TREM2 Gene Dosage Reprograms Microglia Responsivity and Ameliorates Pathological Phenotypes in Alzheimer's Disease Models. Neuron. 2018 Mar 7;97(5):1032-1048.e5. PubMed.
  2. . TREM2 Is a Receptor for β-Amyloid that Mediates Microglial Function. Neuron. 2018 Mar 7;97(5):1023-1031.e7. PubMed.
  3. . TREM2 and Amyloid Beta: A Love-Hate Relationship. Neuron. 2018 Mar 7;97(5):991-993. PubMed.