14 Oct 2024

In neurodegenerative diseases, microglia can help or harm the brain. In the October 3 Journal of Clinical Investigation, scientists led by Ilia Baskakov at the University of Maryland School of Medicine, Baltimore, documented a transition from the former mode to the latter. In mice infected with prion disease, microglia initially scarfed up misfolded prion protein. As they became stuffed with indigestible aggregates, however, their behavior changed. Microglia sidled up to neurons, wrapping their processes around these cells to partially engulf them. The covered neurons did not die, but struggled to function, with many synaptic genes suppressed. At the same time, prion replication took off, and the mice showed the first signs of disease, losing their fear of open spaces.

In mice with symptomatic prion disease, microglia partially envelop many neurons.
This correlates with worsening infection, beginning two weeks before symptoms.
Neuron engulfment may mark a shift from helpful to harmful microglia.

The clinical worsening after microglia envelop neurons suggests the phenomenon is harmful, the authors noted. “With disease progression, microglial behavior changes from positive to negative,” Baskakov told Alzforum. He plans to investigate if preventing this could slow deterioration.

That would be important to know, because the same thing happens in human brain. In postmortem sections from cortical and subcortical brain regions of 57 people with sporadic Creutzfeldt-Jakob disease, the authors found many microglia enveloping neurons. Baskakov believes the finding might apply to other neurodegenerative diseases, as well.

Taking a Bite? In the cortices of mice with prion disease (left), microglia (green) surround neurons (gray). A three-dimensional reconstruction of confocal microscopy images (right) shows prion aggregates (green) in both microglia (red) and enveloped neurons (surface is white). Nuclei are blue. [Courtesy of Makarava et al., 2024.]

Some previous work had suggested that microglia could worsen neurodegenerative disease by chowing down on healthy neurons (for review see Butler et al., 2021). This has been observed in cell culture, where microglia attack tangle-bearing neurons, but not directly in brain (Aug 2018 news; Oct 2021 news).

To find out if it happens in vivo, Baskakov and colleagues infected mice with an aggressive prion known as synthetic strain leading to overweight (SSLOW), which kills mice within four months. The authors previously generated this strain by denaturing recombinant hamster prion protein, injecting it into the brains of wild-type hamsters, then isolating aggregates from those animals and injecting them into healthy ones (Makarava et al., 2010; Jeffrey et al., 2013).

In the cortices of SSLOW-infected, symptomatic mice, first author Natallia Makarava found that microglia had partially encircled up to a third of neurons. Instead of merely extending pseudopodia, as microglia typically do when they eat things, their cell bodies snuggled up to the cells and their processes wrapped the neurons in a bear hug (image above). Very rarely—about 1 percent of the time—microglia fully engulfed the cells. The same partial engulfment occurred in other brain regions, including the hippocampus, striatum, and thalamus. In healthy uninfected mice, by contrast, microglia kept their hands to themselves.

**Shout for Help?** In control mice (top), healthy neurons express abundant GRIN1 (red) on their surface, and microglia (green) keep their distance. In prion-infected mice (bottom), GRIN1 expression falls, and microglia snuggle up. *[Courtesy of Makarava et al., 2024.]*

Examining SSLOW mice at different stages of infection, Makarava and colleagues found that microglia began wrapping up neurons about two weeks before symptoms appeared. This worsened as disease progressed, with almost two-thirds of neurons being surrounded by end-stage disease. In three other models that used less aggressive prion strains, microglia did the same thing, but surrounded fewer neurons. To the authors, this suggested that envelopment correlates with disease stage and severity.

Why does it happen? Possibly, malfunctioning neurons provoke the microglial response. The expression of neuronal genes involved in synaptic transmission, learning, and memory dropped by up to half as disease progressed. In particular, the NMDA receptor subunit GRIN1 dwindled about two weeks before microglia made their move. Perhaps the change signals microglia to finish off struggling cells, the authors suggested (image at left). However, unlike the microglial eating of synapses in neurodegenerative disease, the innate immune complement pathway was not involved in this signaling, the authors found.

In line with the idea that microglia were heading in for the kill, microglia revved up their lysosomal activity as they approached neurons, implying they were preparing for a big meal. Nonetheless, once prion-stuffed microglia contacted a neuron, they seemed to stall out. They did not take up much neuronal material, nor did they kill neurons. Having gorged on misfolded prions, the cells may not be able to consume any more, Baskakov suggested. He noted that an overload of aggregated protein can damage lysosomes. In another sign of this, enveloping microglia began to express apoptotic markers, hinting that they were becoming senescent.

Human Brain, Too. In sections from the hippocampi (left) and occipital cortices (right) of people who died of sporadic CJD, microglia (brown) encircle many neurons (blue, arrows). [Courtesy of Makarava et al., 2024.]

Christina Ising at University Hospital of Cologne, Germany, noted that work from her lab and others has shown that chowing down on tangles can push microglia into senescence (Brelstaff et al., 2021; Karabag et al., 2023). “If these microglia are senescent, it would be very interesting to investigate how prion diseases progress under senolytic treatments at different time points after infection,” she suggested (comment below).

Overall, the data fit a picture where microglia initially help slow disease by mopping up misfolded prions, but eventually sicken and turn harmful, Baskakov said. Prior animal studies support this, showing that microglia improve survival at early stages of prion disease (Zhu et al., 2016; Carroll et al., 2018; Bradford et al., 2022). At later stages, however, eliminating microglia can lengthen lifespan (Gómez-Nicola et al., 2013; Nazmi et al., 2019; Nakagaki et al., 2020).

Next, Baskakov will study the mechanisms underlying this microglia-neuron interaction. He noted that unlike mouse models of Alzheimer’s and similar neurodegenerative diseases, prion-infected mice have the actual disease, suggesting the findings will more directly translate to people.—Madolyn Bowman Rogers

Comments

- Christina Ising
  University Hospital of Koeln
- Posted: 14 Oct 2024
The observation that microglia partially envelop non-apoptotic, viable neurons in prion diseases is quite intriguing. The finding that these enveloped neurons show functional decline might imply that microglia, which amassed high levels of the lysosomal proteins LAMP1 and cathepsin D, were recruited to phagocytose them. But why would they stop halfway? Makarava et al. suggest that this could be due to the development of microglial senescence. While this requires further testing, it is an interesting hypothesis since recent work from Maria Spillantini’s group revealed that microglia phagocytosing tau-bearing neurons develop signs of cellular senescence (Brelstaff et al., 2021). Furthermore, we recently showed that exogenous monomeric tau can induce microglial senescence in culture (Karabag et al., 2023). These mechanisms might also play a role in prion diseases as firstly, microglia enveloping neurons were prion-positive, and secondly, this positivity preceded neuronal envelopment.

However, the mechanistic and clinical relevance of these new findings by Makarava et al. remains elusive. If these microglia are senescent, it would be very interesting to investigate how prion diseases progress under senolytic treatments at different time points after infection. That functional neuronal decline starts ahead of envelopment suggests that, at least initially, microglia may be recruited to protect the brain environment. Methods to follow microglia longitudinally over time in vivo could shed light on the actual sequence of events and show if the body-to-body contact of microglia and neurons in prion diseases is a permanent or reversible state.

Despite there being no evidence to suggest that this phenomenon of (partial) neuronal envelopment by microglia occurs in other neurodegenerative diseases, it is possible that this happens to a small extent in, for example Alzheimer’s disease (AD), as well. But if this phenotype is related to senescence, it will be more difficult to detect and analyze in AD since senescent microglia, while increasingly present, are still low in total number in patient brain samples (Fancy et al., 2024).

References:

Brelstaff JH, Mason M, Katsinelos T, McEwan WA, Ghetti B, Tolkovsky AM, Spillantini MG. Microglia become hypofunctional and release metalloproteases and tau seeds when phagocytosing live neurons with P301S tau aggregates. Sci Adv. 2021 Oct 22;7(43):eabg4980. Epub 2021 Oct 20 PubMed.

Karabag D, Scheiblich H, Griep A, Santarelli F, Schwartz S, Heneka MT, Ising C. Characterizing microglial senescence: Tau as a key player. J Neurochem. 2023 Aug;166(3):517-533. Epub 2023 Jun 5 PubMed.

Fancy NN, Smith AM, Caramello A, Tsartsalis S, Davey K, Muirhead RC, McGarry A, Jenkyns MH, Schneegans E, Chau V, Thomas M, Boulger S, Cheung TK, Adair E, Papageorgopoulou M, Willumsen N, Khozoie C, Gomez-Nicola D, Jackson JS, Matthews PM. Characterisation of premature cell senescence in Alzheimer's disease using single nuclear transcriptomics. Acta Neuropathol. 2024 May 2;147(1):78. PubMed.

View all comments by Christina Ising

- Guy Brown
  University of Cambridge
- Posted: 16 Oct 2024
This paper finds something unexpected during the development of prion disease in mice that may be relevant to Alzheimer’s disease. Microglia are found to partially enwrap neuronal cell bodies, such that up to 40 percent of cortical neurons are covered by microglia. These enwrapped neurons are alive, but show signs of dysfunction. Relatively few neurons are fully enclosed by microglia, i.e., phagocytosed, consistent with no significant neuronal loss in the cortex at the time the mice had to be killed due to weight loss. However, it is unclear whether, had the mice remained alive, the microglia would have gone on to fully phagocytose the neurons, resulting in neuronal death and loss. This form of neuronal death, as a result of inflamed microglia phagocytosing stressed-but-viable neurons, has been proposed for AD (Butler et al., 2021), but microglia have not been caught in the act. In contrast, glial phagocytosis of neurons, “neuronophagia,” has been reported in many acute brain pathologies, such as viral infection. Neuronophagia may not be seen in AD either because it does not occur, or because the rate of neuronal loss is much slower in AD.

The second unexpected finding is that much of the infectious prion is found in the microglia, despite i) microglia not expressing the prion protein, and ii) most of the prion protein being anchored on the surface of neurons or astrocytes. This suggest that microglia are either phagocytosing released prion, prion-infected neuronal parts, or perhaps microglial trogocytosis (nibbling) of the neuronal membrane.

The paper finds that microglial phagocytosis of infectious prion occurs early in the disease, but later these prion-infected microglia enwrap the neurons. This is consistent with microglia attempting to contain the prion infection early on by phagocytosing prion and prion-infected cells, but if the infection is not contained, then this microglial phagocytosis of infected neurons and neuronal parts may contribute to the brain damage.

The authors did not investigate the consequences of the microglial wrapping of neuronal cell bodies, but clearly this could disrupt synapses onto the cell body and thus neuronal networks, potentially disrupting brain function. Those microglia that enwrapped neurons had increased lysosomal markers suggesting they could potentially be phagocytosing synapses or even releasing lysosomal enzymes (by lysosomal exocytosis), which could damage the neurons. The overall findings are reminiscent of AD in that microglia appear to be beneficial early in AD by clearing amyloid, but may be detrimental later on partly by phagocytosis of synapses and neurons. However, the potential roles of microglial wrapping of neurons remains to be explored.

References:

Butler CA, Popescu AS, Kitchener EJ, Allendorf DH, Puigdellívol M, Brown GC. Microglial phagocytosis of neurons in neurodegeneration, and its regulation. J Neurochem. 2021 Aug;158(3):621-639. Epub 2021 Mar 17 PubMed.

View all comments by Guy Brown

- Andrea Tenner
  University of California, Irvine
- Posted: 16 Oct 2024
Engulfment of live neurons and “stressed but viable” neurons has been known for some time (Hugh Perry and Guy Brown come to mind for their papers, reviews and commentary). However, the authors are to be commended for providing critical details of events and their sequence in these processes, importantly demonstrating the contributions not only of the state of the neurons but also the state of the microglia, and how those both change with progression of injury/disease.

The data show initial microglial ingestion of PrPsc is eventually overwhelmed by PrPsc production. The presence of PrPsc or its association with neurons is clearly not sufficient to induce microglia engulfment of the neurons until some unknown trigger polarizes the microglia to engage extensively with the neuron. The authors demonstrated that while there is a decline in levels of neuronal Grin1 prior to microglial engulfment, the apoptotic marker cleaved caspase 3 is not present. The assessment of exposed phosphatidyl serine or annexin 5 would provide support for (or not) a lack of the early known “eat me” signals on neurons as contributing to this process.

While not presented here, one future step would be to perform single cell transcriptomics in this very nice temporally defined system to distinguish between all microglia becoming polarized for aggressive phagocytosis, versus expansions of specific subsets, and exploring the possibility of expanding competing subsets of microglia. This will enable a more effective targeting of the overzealous detrimental microglia while allowing the “reparative” microglia to remain to provide protective effects. Indeed, the authors state “the functional consequences of neuronal envelopment remain unclear.” It will be important to determine, using spatial transcriptomics, whether the microglia (or some of the microglia) “engulfing” but presumably not fully “eating” the neurons, may be supporting neurons or their recovery, so that only the detrimental microglia/microglial functions can be targeted therapeutically in this and similar neurodegenerative disorders. Such subpopulations of microglia have been shown to develop, as recently we published (Schartz et al., 2024).

References:

Schartz ND, Liang HY, Carvalho K, Chu SH, Mendoza-Arvilla A, Petrisko TJ, Gomez-Arboledas A, Mortazavi A, Tenner AJ. C5aR1 antagonism suppresses inflammatory glial responses and alters cellular signaling in an Alzheimer's disease mouse model. Nat Commun. 2024 Aug 15;15(1):7028. PubMed.

View all comments by Andrea Tenner

- Kiran Bhaskar
  University of New Mexico
- Karthikeyan Tangavelou
  New Mexico ADRC, University of New Mexico
- Posted: 16 Oct 2024
This quite interesting manuscript describes the partial enveloping of reactive microglia around viable neurons, which are positive for Pr^PSc and negative for apoptotic markers and decreased neural activity. It is plausible that chronic proteotoxic stress in the Pr^PSc-positive neurons drives apoptosis-resistant senescence, and eventually, the reactive microglia recognize the senescent neurons for phagocytic clearance. In this scenario, microglia may not require the CD11b phagocytic pathway. Despite the increased microglial proliferation and phagocytic activity against the Pr^PSc-viable neurons, Pr^PSc accumulates in the scrapie brain, suggesting that microglial proteostasis is also impaired at later stages of disease progression. This chronic proteotoxic stress can induce senescence in microglia.

Moreover, PrP^Sc-viable neurons may likely secrete pro-survival (of microglia) growth factors, including colony-stimulating factor 1 (CSF1) and interleukin-34 (IL-34), that can induce the proliferation of microglia, causing their replicative senescence as we described in a pending review article (Tangavelou and Bhaskar, 2024). Neuron-microglia contacts in the PrP^Sc condition may induce microglial proliferation, causing replicative senescence or chronic phagocytosis of PrPSc neurons, and impair proteostasis activity, causing proteotoxic stress-driven senescence, leading to accumulation of PrP^Sc in the scrapie brains. This article is fascinating and relevant to understanding the microglia-neuron crosstalk related to prion diseases.

References:

Tangavelou K, Bhaskar K. The Mechanistic Link Between Tau-Driven Proteotoxic Stress and Cellular Senescence in Alzheimer’s Disease. Preprints 2024 Preprint

View all comments by Kiran Bhaskar View all comments by Karthikeyan Tangavelou

- Neil Mabbott
  University of Edinburgh
- Posted: 16 Oct 2024
Makarava and her colleagues reveal an interesting insight into the neuropathological features within the brain during prion disease. They describe how, as the disease progresses, a proportion of the microglia in prion disease-affected brain appear to engulf or partially wrap themselves around neurons. This is a novel finding, and it will, of course, be interesting to learn whether this is a unique property of a subset of the microglia in prion disease-affected brains, or whether similar characteristics are observed in other important neurodegenerative disorders, such as Alzheimer’s disease or Parkinson’s disease.

The study also raises other important questions. The physiological relevance of this novel microglia activity was not determined in this study. Further experiments are now needed to determine whether the engulfment by these microglia affects neuron survival, leading to neurodegeneration. Are the microglia attempting to remove prion-affected neurons? Are the microglia dysregulated in response to prion infection? Conversely, it is also plausible that the microglia are attempting to provide support to prion-infected neurons to prevent their demise as the disease progresses.

No therapies are currently available to treat or cure prion diseases. Addressing these questions may help to reveal novel means to either delay or prevent the development of the neuropathology in brains of patients with prion disease or those with other important neurodegenerative disorders.

View all comments by Neil Mabbott