Because certain variants in TREM2 increase a person's risk of developing Alzheimer’s disease, how this microglial receptor influences AD pathology is an area of intense investigation. Now, researchers led by Zhentao Zhang, Renmin Hospital of Wuhan University, China, say the soluble ectodomain, sTREM2, tempers tau phosphorylation. In the October 21 Nature Communications, they reported that it binds transgelin-2 on the surface of neurons. This, they report, turns off the RhoA-ROCK-GSK3β kinase cascade inside the cells. GSK3β is a major tau kinase. The researchers propose that microglia ramp up sTREM2 secretion to protect neurons and that transgelin-2 is an sTREM2 receptor.

  • Soluble TREM2 shed from microglia binds transgelin-2.
  • GSK3β activity falls, reducing tau phosphorylation.
  • Other scientists question where sTREM2 and transgelin-2 interact.

Other scientists are unconvinced. While some agreed that the sTREM2-transgelin-2 interaction reduced downstream tau pathology, others noted that transgelin-2 contains no transmembrane domain, hence cannot act as a cell-surface receptor. Further, Christian Haass of the German Center for Neurodegenerative Diseases in Munich thinks it is difficult to reconcile the study with TREM2 being a risk factor, since the AD-linked variant H157Y sheds the soluble domain prolifically, leaving less TREM2 on the cell surface to activate microglia. “According to these findings, this mutant should not be a risk variant, but rather a protective variant—how is that explained?” he asked (comment below).

Gary Landreth of the Indiana University School of Medicine, Indianapolis, recalled that TREM2 is no stranger to controversy. “The most prominent question posed about TREM2 is whether it elicits protective or detrimental actions, and the same observational data are used to argue both sides of this issue,” he wrote (comment below).

To explore sTREM2 function, co-first authors Xingyu Zhang and Li Tang added it to cultured neurons from P301S tau mice, which overexpress an FTD-causing mutant tau. Phosphorylation of tau fell at serines 202, 396, 404, and threonine 181. All four residues phosphorylated by GSK3β, hinting that sTREM2 silences the kinase.

The scientists suspected the RhoA-Rho-associated protein kinase (ROCK) cascade might be involved because it stimulates GSK3β and because chipping away at ROCK had been reported to reduce tau pathology in cultured human neurons (Feb 2016 news). When they added sTREM2 to SH-SY5Y neuroblastoma cells the ratio of inactive-to-active RhoA climbed, suggesting that the RhoA-ROCK-GSK3β pathway was silenced.

How might extracellular sTREM2 talk to RhoA within neurons? Zhang and Tang searched for receptors that might bind sTREM2. They isolated the protein and anything that stuck to it from the cell membrane fractions of SH-SY5Y cells using affinity purification, then analyzed the components by mass spectrometry. Among the 17 proteins that came along for the ride, one caught the researchers' eyes: transgelin-2 (TG-2). This 22kDa protein is expressed in most cell types but is particularly abundant in smooth muscle cells, where it binds actin to help it cross-link, hence the word “gel” in transgelin (reviewed by Yin et al., 2019). This was the only protein that bound sTREM2 that might fit the “receptor” category.

Others have plucked TG-2 from cell membranes, and one group claimed it acts as a receptor for metallothionein-2, linking this cysteine-rich protein to the relaxation of smooth muscle cells via inhibition of the RhoA-ROCK pathway (Yin et al., 2018). 

Zhang and colleagues think sTREM2 similarly interacts with TG-2. They report finding sTREM2 bound to transgelin-2 in membrane fractions from wild-type and P301S neurons, and sTREM2 and TG-2 co-localized with neurons, microglia, and astrocytes in hippocampal tissue from P301S mice and from people who had had AD.

TSG12, a small molecule that emerged from a prior screen of TG-2-binding compounds, not only relaxes smooth muscle cells, but Zhang and colleagues found it also quieted GSK3β activity and prevented phosphorylation of tau serines 202 and 396 in neurons from P301S mice. This led the authors to conclude that transgelin-2 acts as a receptor for sTREM2 (image below).

Strange Signaling Microglia release sTREM2, which binds transgelin-2 (TG2) on neurons, according to the authors. This inactivates RhoA, quieting the ROCK-GSK3β pathway and reducing tau phosphorylation. Others reject this model because TG2 has no membrane-spanning domain. [Courtesy of Zhang et al., Nature Communications, 2023.]

Others are doubtful, partly because TG-2 contains no transmembrane domain. “We cannot exclude that it’s membrane-associated, but it’s not membrane-spanning,” Kai Schlepckow, Ludwig Maximilians University, Munich, told Alzforum. Schlepckow accepts that transgelin-2 might mediate effects of sTREM2, but does not think it does so through direct binding at the neuron surface. He believes neuronal receptors for sTREM2 are still to be found.

Michael Willem, also from LMU, agreed with his colleague. He pointed out that neurons express little transgelin-2 compared to other cells, such as smooth muscles. “Much more work needs to be done to uncover a specific sTREM2 receptor selectively expressed by neurons to explain the observed changes [in tau phosphorylation],” he wrote (comment below).

Zhang and Tang report that overexpressing sTREM2 in the hippocampi of P301S mice tempered GSK3β’s activity, lowered p-tau, and preserved synapses (image below). This translated to better memory, since the mice found a hidden platform in a water maze faster than P301S controls, and spent more time exploring the new arm of a Y-maze. These effects ceased when the researchers knocked down transgelin-2 in the hippocampus.

More sTREM2, Less p-Tau. P301S mice (left) have higher phospho-tau levels in the hippocampus than mice overexpressing both P301S and sTREM2 (right). [Courtesy of Zhang et al., Nature Communications, 2023.]

The scientists mimicked the protective effects with a brain-permeable, 13-amino-acid fragment of sTREM2. When they injected it every five days into the abdomens of P301S mice beginning when they were 3 months old, the mice had made half as much hippocampal p-tau, sprouted more synapses, and performed better on the water maze test than did untreated controls. “While Zhang et al. do not address how microglia in the AD brain […] increase the production of sTREM2 and the inhibition of the Rho-ROCK-GSK-3β pathway to prevent tau hyperphosphorylation, they have identified residues 77-89 of sTREM2 as a potential therapeutic peptide to achieve this goal,” wrote Khalid Iqbal of the New York State Institute for Basic Research in Staten Island (comment below).—Chelsea Weidman Burke


  1. We and others have consistently shown that the LOAD-associated TREM2 H157Y variant—which occurs right at TREM2’s ADAM10/17 cleavage site—increases shedding of sTREM2. It therefore decreases TREM2 signaling and function, because signaling-competent cell surface TREM2 is reduced. However, according to the authors’ findings, this mutant should not be a risk variant, but rather a protective variant. How is that explained?

  2. The actions of TREM2 are extraordinarily complicated and not easily understood, at least so far. Some of the essential issues have been confounded by the complexity of the science, the use of a narrow range of disease endpoints, and the use of only a few models, which have so far yielded largely poorly reproducible effects. Mostly, we are trying to unravel mechanisms that we haven’t observed before and don’t really understand. We don’t have an appreciation of the range of the biological activities of the various TREM2 proteoforms. It is controversial what cell types it targets, and what is the nature of the heterogeneity of sTREM2 proteoforms found in the brain. The most prominent question about TREM2 is whether it elicits protective or detrimental actions, and the same observational data are used to argue both sides of this issue.

    The present paper has focused our attention on a question central to understanding exactly how the immune system orchestrates its responses to CNS disease. Do the soluble forms of TREM2 act as a ligand for cell surface receptors on one or more cell types, or does TREM2 also act as a microglial-specific receptor for disease-related ligands? 

    The present study demonstrates that sTREM2 is secreted where it then binds to a novel "receptor" on neurons, Transgelin-2, about which little is known. The authors dissected the downstream signaling pathways, which inhibited tau phosphorylation, effected principally by GSK3β. The activation of TG-2 is then linked to a number of endpoints that are reflective of a "protective" phenotype. Overall, the authors have a reported a number of genuinely remarkable findings which, if verified, may provide new insight into many of the actions elicited by TREM2 and the underlying mechanisms.

  3. The biology of soluble TREM2 (sTREM2), shed from activated microglial cells from its precursor TREM2, is a focus of AD research because TREM2-activating antibodies, which are used in clinical studies, bind and stabilize sTREM2 in the brain of AD mouse models (Schlepckow et al., 2023). We need to understand the physiological function of sTREM2 in this respect.

    I have three criticisms of this paper, which make it difficult for me to agree with the authors’ arguments.

    First, transgelin-2 is by no means a receptor at the cell membrane, and is not expressed in particularly huge amounts in neurons. However, Fig. 4a gives the impression that this is the case; it should be corrected by the authors. I offer an example of such a correction:

    Second, according to transcriptomics data, other cells, such as smooth muscle and endothelial cells, have much higher levels of transgelin-2 (TG2 or TALGN2 gene) as shown below, with lower expression in neurons (2023 ©Brain RNA-Seq), which may vary between specific populations.

    Finally, in the supplement the authors provide more insight into the experimental outcome of the screen with Transgelin-2/sTREM2 binding. A list of other proteins with no association to the cell surface was shown in the supplement. Here, for example, tubulin beta 2A (TUBB2A), a major constituent of microtubules, and the apolipoprotein clusterin (ApoJ) were identified to have even higher specificity than that shown for TG2.

    ApoJ and ApoE are very interesting candidates for sTREM2 binding but the physiology for such a connection was not provided by Zhang et al. In this light I am puzzled about the effects that were shown for sTREM2-treated cells. Much more work needs to be done to uncover a specific sTREM2 receptor selectively expressed by neurons to explain the observed changes.


    . Stimulation of TREM2 with agonistic antibodies-an emerging therapeutic option for Alzheimer's disease. Lancet Neurol. 2023 Nov;22(11):1048-1060. PubMed.

  4. The hallmarks of AD pathology are neurodegeneration associated with tau and Aβ pathologies. Studies in cultured cells and animal models reported over many years have indicated that each one of these three lesions can lead to one of the other two, creating a vicious cycle. While the contribution of each one of these pathologies to the clinical phenotype probably varies from AD subtype to subtype, hyperphosphorylated tau pathology is suspected to play the key role. The AD brain attempts to repair/detoxify itself of each of the three pathologies but, at the end, the loss exceeds the gain. Zhang et al.’s findings that soluble TREM2 ameliorates tau hyperphosphorylation and cognitive deficits through activating transgelin-2 would seem to demonstrate that activation of microglia appears to be one such unsuccessful protective attempt against tau pathology in AD.

    Similarly, we found that AD brain responds to neurodegeneration by activating neurogenesis, which fails to materialize probably because of the lack of sufficient neurotrophic support (Li et al., 2008). Treatment with a neurotrophic ciliary neurotrophic factor-mimetic peptide, P021, which promotes neurogenesis and neural plasticity, inhibits neurodegeneration and both tau and Aβ pathologies, and reverses cognitive impairment in 3x Tg-AD transgenic mouse model of AD (Kazim et al., 2014; Baazaoui et al, 2017). 

    While Zhang et al. do not address how microglia in AD brain can be directed to be activated in a way to increase the production of sTREM2 and the inhibition of the Rho-ROCK-GSK-3β pathway to prevent hyperphosphorylation of tau, they have identified residues 77-89 of sTREM2 as a potential therapeutic peptide to achieve this goal.


    . Failure of neuronal maturation in Alzheimer disease dentate gyrus. J Neuropathol Exp Neurol. 2008 Jan;67(1):78-84. PubMed.

    . Disease modifying effect of chronic oral treatment with a neurotrophic peptidergic compound in a triple transgenic mouse model of Alzheimer's disease. Neurobiol Dis. 2014 Nov;71:110-30. Epub 2014 Jul 15 PubMed.

    . Prevention of dendritic and synaptic deficits and cognitive impairment with a neurotrophic compound. Alzheimers Res Ther. 2017 Jun 27;9(1):45. PubMed.

  5. Another clue is revealed in the mystery of soluble TREM2, unraveling its receptor as well as its function, and surprisingly, in neurons!

    The authors further explored the function of sTREM2 on neurons, which was observed in the 5xFAD mouse model of cerebral amyloidosis expressing the human common variant of TREM2 (Song et al., 2018). They identified transgelin-2 expressed on neurons as the receptor for sTREM2, which is predominantly expressed and shed by microglia in the brain. They reveal a potential mechanism of this neuro-immune interaction by sTREM2 binding to TG2, resulting in deactivation of RhoA-ROCK-GSK3β pathway and thereby less tau phosphorylation. Protective effects of reduced phosphorylated tau were reflected in amelioration of synapse loss and of cognitive deficits in the P301S mouse model of tauopathy.

    The authors identified an active sequence of sTREM2, the amino acid 77-89 peptide, which upon administration into the tauopathy model was responsible for protective effects on tau phosphorylation. The authors did not report whether increased sTREM2 in the P301S mice ameliorated tau-mediated neuronal loss and/or brain atrophy.

    The observations of reduced phosphorylated tau, GSK3β activity, cognitive impairments and synaptic loss track with a study by Jin-Tai Yu and colleagues (Jiang et al., 2016), who showed that overexpression of TREM2 resulted in similar neuroprotective phenotypes in the same P301S mouse model of tauopathy. It is possible that these effects were at least in part due to increased sTREM2 levels in the brain.

    Based on these results, one would speculate that TREM2 variants that reduce sTREM2 shedding or influence receptor-ligand binding would further drive tau hyperphosphorylation, thereby exacerbating associated synaptic and neuronal loss. On the contrary, our lab has reported that P301S mice expressing the human p.R47H TREM2 variant (Gratuze et al., 2020), occurring within the ligand-binding domain of the protein, display reduced phosphorylated tau, which was protective against both synaptic and neuronal loss. Similar studies from P301S/TREM2 knockout mice have reported a protective effect on brain atrophy, attributing it to attenuated microglial reactivity (Leyns et al., 2017; Sayed et al., 2018). 

    Nonetheless, this study is a step toward comprehending the potential roles of sTREM2 on neurons and should be considered while developing treatments aimed at TREM2. Considering the contrasting impact of TREM2 and microglial reactivity on tau and amyloid plaques, it would be intriguing to investigate the potential influence of sTREM2/TG2 interaction on neuritic plaque pathology.


    . TREM2 modifies microglial phenotype and provides neuroprotection in P301S tau transgenic mice. Neuropharmacology. 2016 Jun;105:196-206. Epub 2016 Jan 21 PubMed.

    . Impact of TREM2R47H variant on tau pathology-induced gliosis and neurodegeneration. J Clin Invest. 2020 Sep 1;130(9):4954-4968. 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.

    . Differential effects of partial and complete loss of TREM2 on microglial injury response and tauopathy. Proc Natl Acad Sci U S A. 2018 Oct 2;115(40):10172-10177. Epub 2018 Sep 19 PubMed.

  6. The take-home message of this interesting study is that the RhoA/ROCK/GSK3β pathway appears to be an important regulator of tau phosphorylation, and that this pathway can be modulated by sTREM2. In a therapeutic approach, the authors designed a small peptide mimicking the effects of sTREM2, which may originate from microglia. They also were able to identify effects on tau phosphorylation by inhibition of ROCK using Y-27632, a commonly used ROCK inhibitor. This could open up further avenues to interact with tau phosphorylation.

Make a Comment

To make a comment you must login or register.


Mutation Interactive Images Citations

  1. TREM2

Mutations Citations

  1. TREM2 H157Y

Research Models Citations

  1. Tau P301S (Line PS19)

News Citations

  1. ROCK’n the Tau Pathway?

Paper Citations

  1. . Transgelin-2: Biochemical and Clinical Implications in Cancer and Asthma. Trends Biochem Sci. 2019 Oct;44(10):885-896. Epub 2019 Jun 27 PubMed.
  2. . Transgelin-2 as a therapeutic target for asthmatic pulmonary resistance. Sci Transl Med. 2018 Feb 7;10(427) PubMed.

Further Reading

Primary Papers

  1. . Soluble TREM2 ameliorates tau phosphorylation and cognitive deficits through activating transgelin-2 in Alzheimer's disease. Nat Commun. 2023 Oct 21;14(1):6670. PubMed.