Certain mutations in TREM2 strongly increase risk for late-onset Alzheimer’s. Could tweaking the function of this microglial receptor help people with the disease? A new study suggests as much. In the August 28 Science Translational Medicine, researchers led by Christian Haass at the German Center for Neurodegenerative Diseases in Munich report that people with Alzheimer’s pathology who had more of a soluble TREM2 fragment in their cerebrospinal fluid declined more slowly on memory tests than did those with less sTREM2. Haass had reported preliminary data at the 2019 AD/PD conference in Lisbon, Portugal (Apr 2019 conference news). “Activating TREM2-dependent microglial functions could be a protective therapeutic approach,” Haass told Alzforum.

  • In AD, higher CSF sTREM2 correlates with slower cognitive decline.
  • This holds true regardless of disease stage.
  • Does this mean activating TREM2 may be beneficial?

Most pathological TREM2 variants disrupt the protein’s function (Apr 2017 conference newsJan 2018 news). Since the receptor supports the health of microglia and their responses to amyloid and other brain insults, researchers believe loss of TREM2 function bodes ill for the brain (Aug 2017 news; Sep 2017 news; Mar 2018 news). Nonetheless, depending on disease stage and pathology, some studies have delineated harmful roles for TREM2 (Jul 2018 conference news). 

To better understand the role of TREM2 in disease, Haass and colleagues turned to the Alzheimer’s Disease Neuroimaging Initiative. Joint first authors Michael Ewers, Nicolai Franzmeier, and Marc Suárez-Calvet analyzed data from 385 participants, average age 73. One hundred of them were cognitively normal and free of AD pathology, as judged by CSF levels of Aβ42 and p181-tau. The remaining 285 tested positive for amyloid and p-tau. In this latter group, 35 were cognitively healthy, 184 had mild cognitive impairment, and 66 had dementia. All underwent lumbar puncture at baseline and subsequently took repeated cognitive tests over an average of four years.

Memory Preservation. People with AD pathology whose CSF sTREM/p181-tau ratio was above the group median (blue line) declined more slowly on memory tests than those whose sTREM/p181-tau ratio was below the median (red line). [Courtesy of Ewers et al., Science Translational Medicine.]

As expected from previous studies, baseline CSF sTREM2 was higher in people with AD pathology than in cognitively and biomarker-normal controls. But was the sTREM2 helpful or harmful? To find out, the researchers controlled for not just the known confounders of age, sex, and education, but also for clinical status and CSF Aβ42 and p181-tau. That way, they could compare people at the same disease stage and level of pathology. Their analysis showed that in people with comparable disease, higher baseline sTREM2 associated with slower decline over several years on an episodic memory composite, the ADNI-MEM. When the researchers used the sTREM2/p181-tau ratio, higher baseline numbers correlated with slower decline on both the ADNI-MEM and the ADAS13, a measure of global cognition.

The findings held in both the MCI and dementia subgroups. A Cohen’s d analysis, which measures how one variable affects another, calculated that CSF sTREM2 buoyed the ADNI-MEM score with an effect size of 0.43. This was similar to the effect of p181-tau on cognition—0.47. These are considered small to medium effect sizes. A person’s baseline CSF sTREM2/p181-tau ratio also associated with slower progression from cognitively healthy to MCI and from MCI to AD.

In addition to the cognitive protection, the researchers found hints of a link between sTREM2 and hippocampal preservation. Among 184 participants with CSF AD pathology who had undergone longitudinal MRI imaging, high baseline sTREM2 or high sTREM2/p181-tau was linked with slower hippocampal atrophy, although the results missed statistical significance. A previous cross-sectional study by Haass and colleagues at DZNE and at the Barcelonaβeta Brain Research Center in Spain had correlated high CSF sTREM2 with more gray matter in several cortical regions affected by AD, including temporal and precuneus. This was seen in 27 people with MCI, but not in AD (Gispert et al., 2016). However, it was unclear if the greater volume represented preserved gray matter, or swelling due to inflammation (Aug 2016 conference news). 

“These data argue that sTREM2 has salutary actions on disease pathogenesis,” Gary Landreth at Indiana University School of Medicine in Indianapolis wrote to Alzforum (full comment below). A recent mouse study supports this conclusion, reporting that sTREM2 injected into 5xFAD mouse brain activated microglia to clean up plaques (Apr 2019 news). 

Does this mean that activating TREM2 on the microglial cell surface would help the brain? Haass thinks yes. Because sTREM2 is cleaved from the full-length protein at the cell surface, he believes it likely reflects microglial TREM2 signaling. However, others question this. Landreth pointed to research claiming that up to a quarter of sTREM2 arises from alternative splicing, suggesting it may not correlate with TREM2 signaling (Del-Aguila et al., 2019). Even sTREM2 generated by shedding might oppose signaling if it lowers the amount of intact receptor available on the cell surface, he wrote. David Hansen at Genentech in South San Francisco suggested that sTREM2 shedding could be a protective mechanism that lowers harmful microglial signaling through full-lengthTREM2. In that case, activating TREM2 would be counterproductive for AD patients. Or, sTREM2 might have nothing to do with signaling, and could simply be a marker for microglial abundance. “Relating CSF sTREM2 measurements with additional biomarkers of microglial abundance and active TREM2 signaling will help resolve these uncertainties,” Hansen wrote to Alzforum (full comment below).

Many pharmaceutical companies are developing TREM2 activators. However, Haass cautioned that too much microglial activity can be as bad as too little. His lab recently reported that in progranulin-knockout mice, hyperactive microglia hasten neurodegeneration (Apr 2019 conference news). He believes more basic research is needed before taking TREM2 activators to trials. “We should be very cautious, because nobody knows if microglial activation is reversible,” Haass said. Several groups, including Haass’ lab, have developed TREM2 antibodies, with at least one already in Phase 1 (May 2019 conference news). 

Given that the relationship between sTREM2 and TREM2 signaling remains murky, future trials will need better readouts for the latter, Haass believes. He suggested that changes in microglial lipid metabolism might make a more direct marker for TREM2 activation than CSF sTREM2. TREM2 senses extracellular lipids such as myelin debris and cholesterol. Microglia expressing pathologic mutants of the receptor cannot properly take up or process them, according to research from Denali Therapeutics in South San Francisco (Apr 2019 conference news). This suggests that buildup of lipids such as cholesterol esters could be a marker of poor TREM2 function. In collaboration with Denali, Haass is developing a lipid readout for TREM2 activity.—Madolyn Bowman Rogers


  1. The roles of TREM2 in AD pathogenesis have commanded substantial attention following from the recognition that TREM2 variants are associated with dramatically elevated AD risk. Previous work from this group reported that soluble forms of TREM2 were found at elevated levels in the CSF of AD patients and those with mild cognitive impairment (MCI). Soluble TREM2 (sTREM2) increases were associated with phospho-tau levels in CSF. TREM2 is cleaved at the microglial cell surface by ADAM proteinases, releasing the extracellular soluble domain into the interstitial space, eventually appearing in the CSF. It has been tacitly assumed that release of sTREM2 was reflective of the microglial activation that accompanies disease pathogenesis. Subsequently, elevated CSF sTREM2 was reported to be linked to higher gray-matter volume (Gispert et al., 2016), and disease-associated SNPs at the MS4A locus (Deming et al., 2019). 

    In the present study, Ewers et al. report that subjects who are biomarker-positive (Aβ42 and p-tau) have higher CSF sTREM2 at baseline and exhibit slower disease progression, as evidenced by memory and cognition scores. Indeed, individuals with higher ratios of CSF sTREM2 to phospho-tau were less likely to convert to AD and had slower rates of cognitive loss. These data argue that sTREM2 has salutary actions on disease pathogenesis.

    These findings highlight our ignorance of TREM2 biology in the brain. The concentrations of sTREM2 in the CSF generally have been interpreted as reflective of microglial “activation.” The Ewers study raises the question more broadly of whether increased microglial “activation” favors disease exacerbation or attenuation. The answer is likely to be nuanced.

    First and foremost, there are questions about how sTREM2 is generated. It is clear that the full-length, membrane-associated receptor is shed from the membrane by ADAM10/17, releasing the extracellular domain. It is unclear whether ADAM-mediated cleavage is constitutive or regulated, since ADAM activity is subject to physiological regulation. Existing data suggest that the normal cleavage of TREM2 is quantitatively significant, resulting in reducing overall receptor levels. It also has been postulated that TREM2 cleavage is ligand-stimulated, which would serve to preferentially inactivate the active forms of the receptor. However, the TREM2 gene is alternatively spliced, yielding TREM2 species that do not possess a transmembrane domain and are predicted to be secreted. Del-Aguila et al. reported that up to 25 percent of sTREM2 in the CSF could arise from these transcripts, but this has yet to be demonstrated (Del-Aguila et al., 2019). 

    A provocative study by Zhong and colleagues forces a reevaluation of the significance of sTREM2. These authors report that sTREM2 has cell-autonomous actions through autocrine signaling on microglia whereby the sTREM2 binds to an unknown entity on the microglial cell surface and elicits a broad range of disease-attenuating effects as well as improved behavior (Zhong et al., 2019). 

    Overall, there are more questions than answers.


    . Cerebrospinal fluid sTREM2 levels are associated with gray matter volume increases and reduced diffusivity in early Alzheimer's disease. Alzheimers Dement. 2016 Dec;12(12):1259-1272. Epub 2016 Jul 14 PubMed.

    . The MS4A gene cluster is a key modulator of soluble TREM2 and Alzheimer's disease risk. Sci Transl Med. 2019 Aug 14;11(505) PubMed.

    . TREM2 brain transcript-specific studies in AD and TREM2 mutation carriers. Mol Neurodegener. 2019 May 8;14(1):18. PubMed.

    . Soluble TREM2 ameliorates pathological phenotypes by modulating microglial functions in an Alzheimer's disease model. Nat Commun. 2019 Mar 25;10(1):1365. PubMed.

  2. The findings of Ewers et al. provide a valuable advance for understanding how measurements of the biomarker CSF sTREM2 relate to the rate of disease progression around the early symptomatic stages of Alzheimer’s. What the data imply, as far as TREM2 and microglial activity in AD progression go, is still uncertain because we don’t really know what CSF sTREM2 is telling us biologically. The authors favor the idea that elevated CSF sTREM2 reflects increased TREM2 surface levels and TREM2 signaling activity in brain microglia, which could be true. A second possibility is that CSF sTREM2 is a benign reflection of microglial cell number. In either case, the correlation between elevated sTREM2 and attenuated cognitive decline would suggest that a more robust microglial response to AD pathology would be helpful at this stage (Hamelin et al., 2016; Focke et al., 2018). 

    However, a third possibility is that microglial TREM2 contributes to neurodegeneration at this disease stage (Leyns et al., 2017), and that increased shedding of sTREM2 is a protective mechanism. Relating CSF sTREM2 measurements to biomarkers of microglial abundance and active TREM2 signaling will help resolve these uncertainties.


    . Early and protective microglial activation in Alzheimer's disease: a prospective study using 18F-DPA-714 PET imaging. Brain. 2016 Apr;139(Pt 4):1252-64. Epub 2016 Mar 15 PubMed.

    . Early and longitudinal microglial activation but not amyloid accumulation predict cognitive outcome in PS2APP mice. J Nucl Med. 2018 Sep 27; PubMed.

    . TREM2 deficiency attenuates neuroinflammation and protects against neurodegeneration in a mouse model of tauopathy. Proc Natl Acad Sci U S A. 2017 Oct 24;114(43):11524-11529. Epub 2017 Oct 9 PubMed.

  3. We want to clarify an important point. The goal of our study was not to finally determine if sTREM2 or full-length TREM2 exert a protective function, or which variant needs to be therapeutically modified. Rather, we wanted to find out whether TREM2 is protective in AD patients or not. The only way to do this at the moment is by measuring sTREM2 in CSF and correlating its levels with longitudinal cognitive/clinical scores. Because there are major efforts in the pharmaceutical industry to design therapeutic strategies capable of modulating TREM2 activity, one should better know early on if modulating TREM2 expression/function is beneficial or not.

    It is also important to note that we do not find significant amounts of sTREM2 in CSF derived from alternatively spliced TREM2.  We analyzed a set of human CSF samples using a novel antibody that detects selectively ADAM10-cleaved sTREM2, in comparison with an antibody that detects all forms of sTREM2. This revealed that almost all sTREM2 in CSF is derived by proteolytic shedding from the full-length protein. 

  4. These findings fit quite elegantly with our recent findings and hypothesis that innate immune activity modulates development of presymptomatic AD, and that markers of immune activation are generally beneficial (Meyer et al., 2019). For various reasons we could not include sTREM2 among the CSF immune markers analyzed. A critical feature of both the cited publication and our analyses is that the apparent "effect" of immune activating cytokines and chemokines is not necessarily on progression of AD pathology, but rather on the symptomatic expression of a given degree of such pathology. These "effects" appear to modify not only cross-sectional cognitive performance (ADAS-Cog-11 in our analyses) and four-year cognitive decline, but even clinical diagnosis itself.


    . Hypothesis: cerebrospinal fluid protein markers suggest a pathway toward symptomatic resilience to AD pathology. Alzheimers Dement. 2019 Aug 9; PubMed.

Make a Comment

To make a comment you must login or register.


Mutation Interactive Images Citations

  1. TREM2

News Citations

  1. Parsing How Alzheimer’s Genetic Risk Works Through Microglia
  2. New Evidence Confirms TREM2 Binds Aβ, Drives Protective Response
  3. New Mouse Models Reveal Unexpected Property of TREM2
  4. Without TREM2, Microglia Run Out of Gas
  5. ApoE and Trem2 Flip a Microglial Switch in Neurodegenerative Disease
  6. TREM2 Binds Aβ, Reprograms Microglia to Curb Plaques
  7. TREM2: Diehard Microglial Supporter, Consequences Be DAMed
  8. Refining Models of Amyloid Accumulation in Alzheimer’s Disease
  9. Cut Loose, Soluble TREM2 Beckons Microglia to Mop Up Plaques
  10. Antibodies Against Microglial Receptors TREM2 and CD33 Head to Trials
  11. Could Greasing the Wheels of Lipid Processing Treat Alzheimer’s?

Research Models Citations

  1. 5xFAD (B6SJL)

Paper Citations

  1. . Cerebrospinal fluid sTREM2 levels are associated with gray matter volume increases and reduced diffusivity in early Alzheimer's disease. Alzheimers Dement. 2016 Dec;12(12):1259-1272. Epub 2016 Jul 14 PubMed.
  2. . TREM2 brain transcript-specific studies in AD and TREM2 mutation carriers. Mol Neurodegener. 2019 May 8;14(1):18. PubMed.

Further Reading