Ever since linking loss-of-function mutations in the microglial gene TREM2 to a higher risk of Alzheimer’s disease, scientists have been developing antibodies to boost TREM2 signaling. The latest? A human antibody bestowed with enhanced ability to cross the blood-brain barrier. Kathryn Monroe, Pascal Sanchez, and colleagues at Denali Therapeutics, South San Francisco, report the preclinical characterization of the antibody in the January 12 Nature Neuroscience. It drove TREM2 signaling in cultured human microglia, inducing their proliferation and enhancing their phagocytosis. In a mouse model of amyloidosis, the antibody slid into the brain, increased microglial activity, and boosted brain glucose metabolism. A first Phase 1 trial is ongoing.

  • TREM2 antibody sports transferrin receptor binding domain to ease entry into brain.
  • It activated microglia and boosted brain glucose metabolism.
  • The antibody is in a first Phase 1 trial.

“This is a nifty and elegant approach to get an antibody into the brain by promoting transcytosis,” John Lukens of the University of Virginia, Charlottesville, told Alzforum. Christian Haass of the German Center for Neurodegenerative Diseases in Munich co-authored this work. “The brain shuttle is very important because it allows for dramatically lower antibody concentrations in the periphery, which may decrease the risk of side effects,” Haass wrote to Alzforum.

TREM2 signaling prompts microglia to move, proliferate, and engulf debris such as amyloid plaques. Previously, Haass' group had developed the anti-TREM2 mouse antibody 4D9, which roused microglia to clear amyloid plaques in mice (Mar 2020 news). Haass partnered with Denali to develop a similar antibody against human TREM2 for clinical use that would efficiently slide into the brain. The currently used therapeutic antibodies have a hard time crossing the blood-brain barrier (BBB), such that less than 1 percent of what is infused or injected under the skin gets into the brain.

Co-first authors Bettina van Lengerich, Lihong Zhan, and Dan Xia tapped their brain shuttle program (May 2020 news). It involves hitching a human transferrin receptor binding domain onto cargo of interest, such as an antibody or protein. Transferrin receptors (TfRs) then carry vehicle cum cargo across the BBB. In this case, the scientists inserted the TfR domain, aka antibody transport vehicle (ATV), into the Fc region of 4D9 (see image below). Other companies are trying a similar approach. For example, Roche does it with its brain-shuttle version of gantenerumab (Dec 2021 conference news) and BioArctic collaborates with researchers at Uppsala University, Sweden, on their TfR-based brain transporter technology platform (Hultqvist et al., 2017). 

Subtle Shuttle. Denali’s agonistic anti-human TREM2 antibody sports an ATV (orange) on the end of its stalk region. Leu234Ala mutations (green dots) minimize binding to antibody Fc receptors, which can cause neuroinflammation. [Courtesy of van Lengerich et al., Nature Neuroscience, 2023.]

After intravenous injection into 2- to 3-month-old mice expressing human TfR, sixfold more ATV:4D9 than 4D9 reached the mouse brain. More ATV:4D9 than 4D9 ended up in TfR-rich bone marrow as well, but no immune-cell changes were detected after 12 weekly injections.

ATV:4D9 latched onto microglia that expressed Iba-1 and had stubby branches, two signs of reactivity. One day after ATV:4D9 injection, microglia had shifted their gene expression from homeostatic to reactive, upregulating genes involved in oxidative phosphorylation. One week after treatment, freshly isolated microglia had settled back into their homeostatic state and shape.

Confident that the ATV facilitated brain entry as expected, the scientists generated a human counterpart to 4D9. They injected mice and rats with the extracellular domain of human TREM2, isolated antibodies, and chose the one that best activated TREM2, as measured by increased phosphorylation of the downstream kinase Syk. The researchers then inserted the ATV sequence into the Fc region of this bivalent IgG1, which has a different Fab sequence than 4D9. The resulting antibody, dubbed ATV:TREM2, bound the TREM2 extracellular domain near the ADAM17 protease cleavage site, where 4D9 binds.

ATV:TREM2 doubled TREM2 signaling in cultured human macrophages treated with the lipid phosphatidylserine, a physiological ligand for TREM2, as judged by Syk phosphorylation. Curiously, regardless of its BBB shuttle function, the ATV seemed to boost antibody activity. In cultured human kidney cells, ATV:TREM2 worked better than a version without the ATV, and bivalent antibodies outdid their monovalent equivalents.

Even though the 1.4 μM affinity of ATV:TREM2 for TfR dimers on the cell surface pales in comparison to the 2 nM affinity for TREM2, the authors believe that TfR binding helps cluster TREM2, which is known to amplify signaling (see image below). In essence, the ATV turns a bivalent antibody into a tetravalent one. An anti-TREM2 tetravalent antibody was recently reported to boost TREM2 signaling 100-fold (Sep 2022 news).

The More, The Merrier. The monovalent TREM2 antibody (anti-TREM2 MV) did not prompt TREM2 (green) to phosphorylate Syk. The bivalent (anti-TREM2) and ATV:TREM2 monovalent antibody activated TREM2 equally. Surprisingly, bivalent ATV:TREM2 boosted TREM2 signaling even more. [Courtesy of van Lengerich et al., Nature Neuroscience, 2023.]

Indeed, in kidney cells expressing a TREM2/mini biotin ligase chimera in addition to normal TREM2, ATV:TREM2 induced more TREM2 biotinylation than did the antibody sans ATV, suggesting the ATV improved clustering. Immunoprecipitation of TREM2 from the treated cells pulled out TfR, suggesting the two proteins had formed a complex.

Monroe and Sanchez were surprised that the ATV enhanced TREM2 signaling. So was Nimansha Jain of Washington University in St. Louis. “It is very intriguing that the ATV:TREM2-activating antibody enhances the brain exposure and pharmacodynamic microglial responses compared to an anti-TREM2 antibody,” she wrote (comment below). “Moving forward, it will be interesting to see how this increased brain exposure affects Aβ pathology and tau pathology in various mouse models.”

To see how it performed in vivo, the researchers injected ATV:TREM2 into the veins of transgenic mice expressing human TREM2 and human TfR. Compared to the bare TREM2 antibody, more ATV:TREM2 entered the brain. Not only that, it permeated the entire parenchyma, as seen on in vivo SPECT imaging of antibodies labeled with indium-111 (image below).

Accessing the Brain. As seen by SPECT imaging, much more ATV:TREM2 flooded the mouse brain (right) after intravenous injection than anti-TREM2 without the antibody transport vehicle (left). [Courtesy of van Lengerich et al., Nature Neuroscience, 2023.]

ATV:TREM2 increased the number of proliferating microglia as measured by uptake of the modified nucleoside EdU. When taken out of the mice, these microglia engulfed more fluorescently labeled Aβ and myelin than microglia from mice given a control antibody. They also upregulated genes involved in oxidative phosphorylation and glycolysis. Likewise, cultured human microglia given ATV:TREM2 ramped up glucose metabolism.

How would a mouse model of amyloidosis respond to ATV:TREM2? The scientists crossed mice expressing human TREM2/TfR with the 5xFAD line, then injected the antibody into 4.5-month-old offspring. At this age, the mice have widespread amyloid plaques and gliosis. Up to eight days after injection, antibody-treated mice had higher TSPO- and FDG-PET signals in their cortices than did untreated mice, indicating more active microglia and higher brain glucose metabolism, respectively.

Revving Up Microglia, Metabolism. In 5xFAD mice, microglial activity (top) and brain glucose metabolism (bottom) rose one (left), four (middle), and eight (right) days after intravenous injection of ATV:TREM2. [Courtesy of van Lengerich et al., Nature Neuroscience, 2023.]

The scientists did not report whether ATV:TREM2 cleared plaques. “The question remains whether [antibody-induced changes in microglia are] enough to modify disease in any way,” wrote Renzo Mancuso, VIB-Center for Molecular Neurology, Belgium (comment below).

Lukens noted that the antibody’s effects on tauopathy would be critical, because a TREM2 agonistic antibody was recently shown to worsen tangle load in mice (Oct 2022 news). “We are only scratching the surface of how agonist [TREM2] antibodies may be customized for CNS applications,” wrote Chris Bennett, University of Pennsylvania, Philadelphia (comment below). Sanchez told Alzforum that they plan to test the antibody in a dual amyloid/tau model using the APP-SAA knock-in being developed at Jackson labs.

Denali partners with Takeda to test the antibody, now called DNL919, in a Phase 1 study of 80 healthy people in the Netherlands. Denali's Laura Hansen told Alzforum that results should be in by the end of the year. Sanchez is also interested in exploring whether ATV:TREM2 has additive of synergistic effects with anti-amyloid antibody treatments.—Chelsea Weidman Burke


  1. It is very intriguing that the ATV:TREM2-activating antibody enhances brain exposure and pharmacodynamic microglial responses compared to an anti-TREM2 antibody. Moving forward, it will be interesting to see how this increased brain exposure affects Aβ and tau pathology in various mouse models. Additional experiments will lead to a better understanding of the interplay between TREM2 activation, timing of treatment, and effective doses on in vivo changes in AD pathology.

  2. I find this a very stimulating paper that shows the vast potential of collaborations between academics and industry. I think this is the way forward and we should foster these types of interactions.

    The authors describe a robust strategy to induce TREM2 activation in vivo, in mice. It is clear that the strategy works and that the biologicals developed by Denali do increase the biodistribution in the brain and therefore the activity of the TREM2-activating antibodies.

    It is also nice to see multiple indicators of target engagement in the form of transcriptomics and metabolic changes, some of which could be translated to humans using imaging techniques. However, there are a number of questions:

    Why is the activation of TREM2 via antibodies not able to elicit a DAM response? It is certain that TREM2 is necessary for microglia to develop a DAM response, but is it the only receptor engaged by plaques, or are there other receptors that need to be co-activated to trigger a DAM response? Or could this be explained by the specific effect of this antibody, whereas others would have potential to induce a full response?

    Clearly, there is target engagement, because the treatment with the antibody induces several changes in the microglia. Is this enough to modify disease in any way? This paper still misses a key element, which is whether exogenous activation of TREM2 can modify microglial function in such a way that they at least react more robustly—and hopefully clear Aβ.

    Lastly, given recent data that attenuation of the microglial response might be beneficial in tauopathies, it will be essential to determine the effect of this antibody beyond amyloid models of AD.

  3. Human genetics data strongly argue that microglial dysfunction contributes to Alzheimer’s disease. One striking example is the identification of loss-of-function TREM2 mutations that confer greatly increased risk for AD. Meanwhile, a body of murine studies argues that deficient TREM2 function limits protective microglial responses to amyloid pathology such as engulfment and plaque compaction, potentially by impairing appropriate metabolic adaptation.

    Accordingly, multiple groups have tested TREM2 agonist antibodies in animal models and humans as a strategy to treat AD. Studies support the conclusion that enhancing TREM2 agonism reduces pathology and improves behavioral deficits in amyloid models, correlated with microglial proliferation, plaque containment, metabolic fitness, and phagocytosis. Studies vary, however, in which specific disease phenotypes respond to TREM2 agonism. For example, some studies note reduced plaque burden, while others do not. Further, a recent study found that one TREM2 agonist antibody exacerbated tau pathology, while genetic manipulation of TREM2 leads to sometimes conflicting, opposite results between labs and models. In sum, the role of TREM2 agonism in AD treatment is very promising, but our still unclear understanding of when and how it influences AD pathogenesis limits therapy development. A related uncertainty is the degree to which differences between studies are explained by specific TREM2 antibodies used, and their properties.

    This work of van Lengerich and colleagues is exciting for several reasons. First, it introduces a new candidate therapeutic for study in humans with an engineered transferrin receptor binding site, and a defined binding epitope in the stalk region near the ADAM17 cleavage site, similar to a well-characterized murine equivalent. Second, the study builds upon prior understanding of how antibody-based TREM2 agonism affects microglia and the brain, showing that agonism elicits transient microglial states distinct from the well-described DAM phenotype, correlated with many effects that include enhanced proliferation, phagocytosis, glucose metabolism, and mitochondrial metabolism. Third, the authors present evidence that the transferrin receptor binding not only increases delivery to the brain as intended, but also increases TREM2 clustering and signaling via additional interactions with the transferrin receptor, though the degree to which this occurs in vivo remains to be studied.

    These findings enhance our framework for how to measure and describe the effects of TREM2 agonism, and suggest that we are only scratching the surface of how agonist antibodies may be customized for CNS applications. With careful engineering, it may be possible to further tune TREM2 signaling, both quantitatively and qualitatively, to gain even more precise control over microglial responses. It would be interesting to benchmark various TREM2 agonist antibodies across the assays presented in this study, along with murine amyloid models, to better understand which design features are most critical for which downstream effects. 

    By demonstrating new ways to enhance TREM2 signaling, this study raises another question: How much agonism is too much? It seems possible that with too much agonism, microglia may be at risk of adopting detrimental states, and perhaps there are qualitative and quantitative “sweet spots” for optimal therapeutic effects. This interesting study sustains excitement over targeting microglia to treat AD.

  4. In this exciting study, Lengerich and Zhan et al. have grafted the transferrin receptor (TfR) binding site onto two anti-TREM2-stalk-region antibodies to boost the antibody delivery into the CNS in an amyloid mouse model. In line with the previous work with anti-TREM2 treatments, this paper further demonstrates the efficacy of TREM2 agonist antibodies to enhance microglial response against Alzheimer’s pathology through TREM2-dependent signaling. The paper contains several novel and interesting points:

    1. It provides evidence that the “shuttle” approach can be effective in inducing microglia activation using low doses of antibody. This is relevant for therapy, as it may reduce the therapeutic costs and may potentially lead to administration routes different from intravenous infusion, such as subcutaneous administration.
    2. It provides detailed information on a new anti-human TREM2 antibody.

    We are happy to see that the comprehensive characterizations of a single injection of anti-TREM2 antibody in this study recapitulates the induction of microglia proliferation and state shifts induced by a single acute treatment that we reported in previous studies with different antibodies (Ellwanger et al., 2021; Wang et al., 2020). It is interesting that the authors further examine the duration of anti-TREM2-stimulated responses. The new data demonstrates waning of microglia activation 14 days after treatment. It is possible that different antibodies with different pharmacokinetics may have distinct duration. It will be important to follow up these studies with long-term treatments to check the duration of TREM2 activation after multiple injections.

    It is nice to see that the gain-of-function experiments with the anti-TREM2-TfR antibody presented in this study corroborate our previous work on TREM2 loss-of-function with regard to the impact of susceptibility to CSF1 deprivation and mTOR activation (Ulland et al.., 2017; Wang et al., 2015). With two different anti-TREM2 antibodies (all binding to stalk region of TREM2), Lengerich and Zhan et al. showed a promising enhancement of the same pathways. We recently showed that TREM2 signals through DAP12 but also DAP10, at least in an animal model (Wang et al., 2022). Thus, it would be interesting to know whether anti-TREM2 treatment affects both downstream pathways. Another interesting question is whether this enhancement can be applied to some clinically relevant variants such as TREM2 R47H and TREM2 R62H.

    It has been hypothesized that the beneficial effect of anti-TREM2 may depend on timing. Even though many groups have shown the protective role of TREM2 signaling in different amyloid models, recent studies on tau models have raised some concern that TREM2 activation may be detrimental at late-stage disease (Jain et al., 2023; Sayed et al., 2018). It will be important to investigate what the outcome of the TfR-shuttle delivery method is in these models.

    Another question for future studies concerns the potential side effects of global changes in the biodistribution due to TfR-anti-TREM2 fusion. The authors showed no changes in immune cells in different organs. However, the study was limited to single dose in a 12-week period; moreover, no disease models have been investigated. For example, TREM2 blockade has a beneficial impact on tumorigenesis (Molgora et al., 2020). Will enhanced biodistribution of TfR-fused anti-TREM2 antibody to peripheral organs modify the risk for other diseases?

    Finally, a short-term treatment with TfR-fused anti-TREM2 antibody had no impact on Aβ plaques, which is consistent with our previous studies. However, an anti-mouse TREM2 mAb from the same group was shown to reduce plaques after short treatment. It will be important to clarify whether acute and chronic treatment with anti-TREM2 affect Aβ load.

  5. This is an interesting paper. Delivery of large-molecule drugs across the blood-brain barrier (BBB) is increasingly being seen as an achievable goal. Several technologies have been described where following peripheral administration, the molecules can be detected in the brain. The authors have clearly demonstrated this in 2- to 3-month-old mice expressing human TfR, After intravenous injection, ATV-4D9 shows sixfold more permeability than 4D9 across the brain and no immune-cell changes were detected up to 12 weeks after treatment. Binding of 4D9 antibody to transferrin receptor activates microglia and promotes glucose metabolism, it would be interesting to investigate the effectiveness of this antibody conjugated with transferrin on various conformational structures of Aβ, tau, and secretases involved in APP pathways.

    Previously, Genentec has used various bispecific antibodies with optimized binding to the transferrin receptor (TfR) that target β-secretase (BACE1), can cross the BBB, and reduce brain Aβ in human TfR knock-in mice in a TfR affinity-dependent fashion. Intravenous dosing of monkeys with anti-TfR/BACE1 antibodies also reduced Aβ both in cerebral spinal fluid and in brain tissue, and the degree of reduction correlated with the brain concentration of anti-TfR/BACE1 antibody.

Make a Comment

To make a comment you must login or register.


Mutation Interactive Images Citations

  1. TREM2

News Citations

  1. Paper Alert: Mouse TREM2 Antibody Boosts Microglial Plaque Clean-Up
  2. Molecular Transport Vehicle Shuttles Therapies into Brain
  3. Brain Shuttle Could Halve Amount of Gantenerumab Needed
  4. Potent TREM2 Antibody Stirs Microglia to Prune Plaques in Mice
  5. In Mice, TREM2 Antibody Mobilizes Microglia, Yet Worsens Tangles

Research Models Citations

  1. 5xFAD (B6SJL)

Paper Citations

  1. . Bivalent Brain Shuttle Increases Antibody Uptake by Monovalent Binding to the Transferrin Receptor. Theranostics. 2017;7(2):308-318. PubMed.

Other Citations

  1. APP-SAA knock-in

External Citations

  1. study 

Further Reading

Primary Papers

  1. . A TREM2-activating antibody with a blood-brain barrier transport vehicle enhances microglial metabolism in Alzheimer's disease models. Nat Neurosci. 2023 Mar;26(3):416-429. Epub 2023 Jan 12 PubMed.