In Frontotemporal Dementia Model, Do Riled-Up Microglia Help?
Quick Links
Microglia must be finely tuned to benefit the brain, and pushing them too far in either direction can be harmful. In Alzheimer’s disease, these cells can become stuck in a homeostatic state, unable to respond to pathology, while in frontotemporal dementia caused by progranulin mutations, they go wild, wreaking havoc. Or so it seemed. In the January 12 EMBO Journal, researchers led by Christian Haass at Ludwig-Maximilians University in Munich now question this latter idea, showing that hyperactivated microglia may still protect the brain. In mice lacking progranulin, deleting TREM2 calmed neuroinflammation, but nonetheless lysosomal function and synapse loss worsened.
- Deleting TREM2 calms hyperactive microglia in progranulin knockouts.
- Despite this, neurodegeneration worsens.
- This hints that hyperactive microglia may still have some protective function.
“We think lysosomal damage comes first, and the microglial hyperactivity is a response,” Haass told Alzforum.
In other words, microglia rev up in an attempt to protect the brain from the consequences of malfunctioning lysosomes. If so, the best therapeutic strategy for this form of FTD might be to tackle the root problem by bolstering lysosome activity, he suggested.
Philip Van Damme at KU Leuven, Belgium, said the study shows how complex it could be to modulate microglia effectively. “A better understanding of the neuroprotective versus neurotoxic properties of microglia is instrumental for designing therapeutic strategies. Microglial dysfunction seems context-, disease-, and disease-stage dependent,” he wrote to Alzforum.
Inflammation Dimmer. In progranulin knockout mouse brains (top), microglia “heat up” (red), whereas in TREM2 knockouts (middle), they “cool down” (blue). In double knockouts (bottom), microglial activity is similar to wild-type levels (gray). [Courtesy of Reifschneider et al., The EMBO Journal.]
Previously, Haass’ group had reported that losing progranulin supercharged microglia, boosting their appetite and causing them to spew cytokines, potentially damaging the brain (Apr 2019 conference news). This phenotype is the opposite of microglia losing TREM2, which makes them quiescent (May 2017 news).
To test whether hyperactivated microglia contribute to the damage in the FTD brain, Haass, working with Anja Capell and Dominik Paquet at Ludwig-Maximilians, quieted the cells by deleting the master microglial regulator TREM2. First author Anika Reifschneider crossed GRN mice, which lack the progranulin gene, with TREM2 knockouts. As expected, this doused the microglial hyperactivity of GRN knockouts back to wild-type levels, as measured by TSPO PET imaging (see image at right). Likewise, the disease-associated microglia (DAM) gene-expression signature in GRN knockouts was nearly restored to a homeostatic state by TREM2 deletion.
What effect did this dampening of inflammation have on the mice? To the scientists’ surprise, the double knockouts fared worse. Their presynaptic markers synaptophysin and VGAT were lower than in GRN and TREM2 single knockouts, suggesting more lost synapses. Neurodegeneration appeared more severe. At 14 months of age, neurofilament light chain (NfL) in cerebrospinal fluid was three times higher in double knockout mice than in GRN knockouts.
Brain glucose metabolism, another measure of brain health, was as low in double knockouts as GRN knockouts. Double knockouts also had the same lysosomal deficits as GRN knockouts, with the waste product lipofuscin accumulating despite overexpression of many lysosomal genes (Jul 2017 news; Sep 2020 news).
To find out why these problems were as bad or worse in double knockouts as GRN knockouts, the authors compared their brain gene-expression profiles. They found few differences. In particular, there was no increase in neuroinflammatory or synaptic pruning genes in the double knockouts, suggesting no specific triggering of these pathways. Instead, Haass believes the hyperactivated microglia in GRN knockouts provide some protection against neurodegeneration, which is lost upon TREM2 deletion.
Oleg Butovsky at Brigham and Women’s Hospital, Boston, was skeptical of this interpretation, noting that lipid dysfunction and synaptic health may simply be independent of TREM2 and microglial activation. One way to test the theory that microglia help progranulin-deficient mice would be to further activate these cells via antibodies and see if that improves brain function, he suggested.
Blocking Antibodies. Newly generated antibodies (magenta) bind at TREM2’s Ig loop, blocking intracellular signaling. [Courtesy of Reifschneider et al., The EMBO Journal.]
Do the mouse findings apply to human cells? Haass and colleagues investigated using macrophages taken from patients with pathogenic GRN mutations, as well as iPSC-derived microglia lacking GRN. Co-author Kathryn Monroe and colleagues at Denali Therapeutics, South San Francisco, generated two different antibodies that blocked intracellular TREM2 signaling and increased shedding of its soluble fragment (see image above). For both cell types, the antibodies mirrored the effect of TREM2 deletion in mice microglia. Gene expression returned to a nearly homeostatic signature, but lysosomal dysfunction remained.
Haass noted that these antibodies will be useful tools, though they are unsuitable for treating progranulin-deficient FTD. Most antibodies in development activate the receptor (Mar 2020 news; Jun 2020 news).
By having a means to dial down TREM2 signaling, scientists can now address new questions. In particular, Haass wants to compare the gene-expression profiles of microglia treated with agonist versus antagonist antibodies to home in on which genes define beneficial microglial activation. “That would be eye-opening,” he told Alzforum.
Haass also found it encouraging that microglia treated with the antagonistic antibodies could return to a homeostatic state. Previously, he had feared that activating microglia with TREM2 antibodies might trap them in a super-activated state that could damage the brain. This could have rendered TREM2 too risky a target. “Now we know microglia can switch back and forth. That is good news for antibody treatments,” Haass said.—Madolyn Bowman Rogers
References
News Citations
- Parsing How Alzheimer’s Genetic Risk Works Through Microglia
- Paper Alert: TREM2 Crucial for Microglial Activation
- TMEM106B and Progranulin Duke It Out at the Lysosome
- In ALS and FTD, Two Different Routes to TDP-43 Aggregation
- Paper Alert: Mouse TREM2 Antibody Boosts Microglial Plaque Clean-Up
- In Mice, Activating TREM2 Tempers Plaque Toxicity, not Load
Further Reading
Primary Papers
- Reifschneider A, Robinson S, van Lengerich B, Gnörich J, Logan T, Heindl S, Vogt MA, Weidinger E, Riedl L, Wind K, Zatcepin A, Pesämaa I, Haberl S, Nuscher B, Kleinberger G, Klimmt J, Götzl JK, Liesz A, Bürger K, Brendel M, Levin J, Diehl-Schmid J, Suh J, Di Paolo G, Lewcock JW, Monroe KM, Paquet D, Capell A, Haass C. Loss of TREM2 rescues hyperactivation of microglia, but not lysosomal deficits and neurotoxicity in models of progranulin deficiency. EMBO J. 2022 Feb 15;41(4):e109108. Epub 2022 Jan 12 PubMed.
Annotate
To make an annotation you must Login or Register.
Comments
K.U. Leuven and VIB
This interesting study shows that interfering with microglia seems more complicated than anticipated. A better understanding of the neuroprotective versus neurotoxic properties of microglia is instrumental for designing therapeutic strategies. Microglial dysfunction seems context-, disease-, and disease-stage dependent.
Simply reducing the activation of microglia may be too simplistic as an approach, at least for some diseases. It would be of interest to better understand the negative consequences of TREM2 inhibition and TREM2 deletion on lysosomal dysfunction and turnover and accumulation of insoluble TDP-43. In addition, it may be of interest to investigate the consequences of agonistic TREM2 antibodies, as well.
Goizueta Institute @ Emory Brain Health
I have written extensively about the challenge of harnessing the immune system for therapeutic benefit in neurodegenerative diseases (see for example (Golde et al., 2018).
The Goldilock's principle no doubt applies here, as well. Depending on the readout, it's hard to get the just the right temperature (which, in this case, is the immune activation state). Further, what the readout is in preclinical models that is needed to possibly predict clinical impact is debatable.
This elegant paper shows another example of the delicate balance between beneficial and harmful effects of immune activation in the CNS. We have numerous examples where manipulating immune activation states beneficially impacts one pathology, accelerates another and, in some cases, can cause outright neurodegeneration.
There is immense interest in therapeutically manipulating TREM2 and other immune targets in the brain, especially in AD. Indeed, we are already there, testing a few such therapies in humans. In the past, such studies, e.g. of NSAIDs, have probably done as much harm as good, and I would argue that our understanding of potential beneficial and untoward impacts is so nascent that we need a much more cautious data-driven approach.
I would also argue that we need to be be thinking about the peripheral impacts of these manipulations as well. In this case, do TREM2 antagonist antibodies impact osteoclast function and bone mineralization (or other aspects of peripheral immune function)?
I think a Pollyanna-ish view of these emerging therapeutic strategies will likely result in a suboptimal or even untoward outcome. All involved in these types of studies need to engage a bit more to discuss what kind of preclinical data is needed to move a therapeutic strategy forward toward humans. For example, should we insist that if AD is the disease of interest, we test in both amyloid and tau models, and what do we do if a beneficial effect is seen on one pathology and a harmful effect on the other?
References:
Golde TE. Harnessing Immunoproteostasis to Treat Neurodegenerative Disorders. Neuron. 2019 Mar 20;101(6):1003-1015. PubMed.
Make a Comment
To make a comment you must login or register.