As the brain’s resident immune cells, microglia gobble up dangerous gunk. However, according to a March 1 paper in Molecular Neurodegeneration, the cells may sometimes do more harm than good, at least in the earliest stages of Alzheimer’s disease. Researchers led by Charles Glabe at the University of California, Irvine, reported that using the drug PLX3397 to selectively kill off microglia in an AD mouse model reduced accumulation of Aβ inside neurons, and prevented the formation of neuritic plaques. The treated animals also dodged some of the behavioral deficits documented for this mouse line. The researchers propose a cascade in which microglia somehow promote the toxic accumulation of intraneuronal Aβ, which ultimately kills neurons and triggers the formation of neuritic plaques.

  • Treating 5xFAD mice with PLX3397, a CSF1R inhibitor that ablates microglia, prevented neuritic plaque formation.
  • This lessened accumulation of Aβ inside cells.
  • Treated mice performed better on a memory test.

Glabe told Alzforum that he sought to understand the role microglia play in the development of neuritic plaques. Intermeshed with remnants of dead neurons and surrounded by swollen, dystrophic neurites, this form of plaque likely originates from what were once intracellular aggregates of Aβ, Glabe believes. A distinctive and, to Glabe’s mind, suspicious, halo of microglia surrounds neuritic plaques, prompting him to probe whether the cells promoted this pathology or were merely looky-loos.

To address this question, first author Justyna Sosna and colleagues ablated microglia in 5xFAD mice using PLX3397. This small molecule inhibits colony-stimulating factor 1 receptor (CSF1R), which microglia need to survive. The drug drastically reduces numbers of microglia, circulating monocytes, and resident macrophages in other organs; it has been approved for human use to treat some cancers.

Previous work from Kim Green’s lab, also at UC Irvine, found that microglial populations rapidly replenished following ablation. In Green’s lab, treating 10-month-old, plaque-ridden 5xFAD mice with PLX3397 did not reduce plaque burden, yet protected synapses (Apr 2014 newsSpangenberg et al., 2016). 

Sosna asked what would happen if the mice took the drug earlier, at a time when neurons harbor Aβ aggregates but few neuritic plaques have formed yet. The researchers therefore started treating two-month-old mice and continued for three months. At five months, immunofluorescence of brain sections from untreated animals showed widespread Aβ deposition, the vast majority in the form of neuritic plaques. In contrast, animals fed with PLX3397-laced chow had up to 80 percent fewer microglia and 90 percent fewer neuritic plaques throughout the brain, including hippocampus, cortex, and amygdala. The few neuritic plaques that did exist were smaller and more compact than those in control mice, and remaining microglia surrounded them.

The microglial ablation also drastically reduced accumulation of intraneuronal Aβ, to a similar degree as it did neuritic plaques.

No Microglia, No Plaques.

Five-month-old 5xFAD mice had extensive plaques (6E10, green) in the brain (top: coronal section; 1, hippocampus; 2, cortex; 3, thalamus; 4, amygdala). PLX3397 (right panels) dramatically reduced plaques. [Courtesy of Sosna et al., Molecular Neurodegeneration, 2018.]

Using conformation-dependent antibodies that distinguish different types of Aβ oligomer, the researchers investigated how PLX3397 influenced these soluble Aβ species. They found that compared with wild-type mice, five-month-old 5xFADs had more so-called “fibrillar oligomers” in the hippocampus and cortex; these tend to be stable and contain parallel β-sheet structures. In the plasma of 5xFAD animals, the scientists found an abundance of “prefibrillar oligomers,” which are less stable and have anti-parallel β-sheet structure. PLX3397 substantially reduced levels of both of these species of Aβ.

Would this ward off memory problems? The researchers conducted two tests of hippocampal memory: the Y-maze test of spatial memory, and the fear-conditioning test of contextual memory. Five-month-old 5xFAD mice had slight deficits in both relative to wild-type. PLX3397-treated 5xFAD mice performed at wild-type levels on the contextual-fear-memory test, but showed no benefit on the spatial-memory test. Glabe told Alzforum that perhaps the treatment had no extensive benefits because at just five months of age, memory problems in 5xFAD mice are subtle.

The findings point to a role for microglia in the formation of neuritic plaques, Glabe said. Given Green’s previous finding that removing microglia in older, plaque-ridden mice had no effect on Aβ pathology, Glabe’s study suggests that microglia play a formative role early on but not later.

Glabe was surprised that microglia seemed to promote the accumulation of intraneuronal Aβ, and he wasn’t alone. “These results are interesting because they provide an additional line of evidence for the beneficial impact of targeting microglial numbers in AD, but also a bit puzzling because, after all, it is difficult to understand how Aβ production could have been ‘switched-off’ in a transgenic model where the promoters are not microglia or CSF1R-dependent,” commented Diego Gomez-Nicola of the University of Southampton, England. “Data analyzing the levels of APP expression and/or protein levels of Aβ could have clarified this issue, alongside analysis of the proteins involved in APP processing,” he added.

Glabe proposed that microglia might send signals to neurons that coax them into storing intraneuronal Aβ, which ultimately reaches a toxic threshold that kills the neurons; this would create fodder for future neuritic plaques. Future studies should identify those signals, he said.

Alternatively, CSF1R signaling on neurons might promote intraneuronal Aβ accumulation, Glabe offered. A previous study reported that some neurons express the receptor, and that it provides neuroprotective signals (Luo et al., 2013). 

Gary Landreth of Indiana University School of Medicine in Indianapolis also raised this issue. “One of the difficulties with the manuscript arises from the use of PLX3397 to ablate microglia, since it inhibits CSF1R/kit found both on microglia and in a subset of neurons,” he wrote. “Thus, it is not possible to determine if the observed effects on pathology arise from salutary effects of the drug on neurons or from the loss of microglia in the brain.” He added that a careful counting of neuronal numbers would have helped clarify this conundrum.

Landreth also wondered about the nature of the plaque-associated microglia that somehow resisted ablation, speculating that these CSF1R-independent survivors could be endowed with beneficial properties.

“Overall this paper highlights some exciting ideas; however, it falls short on the delivery of clear results implicating microglia in the observed phenotypes and behavioral improvement,” he wrote.

For PLX3397 studies in the brain, mixed results are par for the course. A recent study found that PLX3397 treatment exacerbated neuronal loss in a Parkinson’s mouse model, while a paper submitted to BioRxiv reported the drug was beneficial in a model of epilepsy (Yang et al., 2018; Srivastava et al., 2017). 

The explanation for these contradictory findings may lie in the growing realization that the role of microglia may change depending on disease stage. Researchers increasingly acknowledge that they need to learn to characterize context-dependent microglial phenotypes, and as they do so, expect to discover a variety of distinct microglial populations. Consider yet another new study, in a model of amyotrophic lateral sclerosis. In inducible TDP-43 mice, microglia stood idly by while pathology ramped up, but when TDP-43 expression was shut off, the cells rose up to expand, change gene expression, and prevent neurodegeneration. Giving PLX3397 during this “recovery phase” blocked the mice from regaining motor function (Mar 2018 news).—Jessica Shugart

Comments

  1. This is an interesting study, expanding recent literature on the therapeutic potential of CSF1R targeting in AD (Olmos-Alonso, 2016Dagher et al., 2015Spangenberg et al., 2016). 

    Here, a preventative treatment with a CSF1R inhibitor was applied, in order to deplete microglia before the formation of neuritic plaques. Surprisingly, depletion of microglia caused an unexpected impairment of amyloid levels, leading to a mild improvement in some behavioral tests. These results are interesting because they provide an additional line of evidence for the beneficial impact of targeting microglial numbers in AD, but also a bit puzzling because, after all, it is difficult to understand how Aβ production could have been “switched-off” in a transgenic model where the promoters are not microglia or CSF1R-dependent. Data analyzing the levels of APP expression and/or protein levels of Aβ could have clarified this issue, alongside analysis of the proteins involved in APP processing.

    Also, if the intervention was indeed so effective in arresting the formation of amyloid pathology, one would have expected a complete prevention of the behavioral deficits (as this is a preventative approach), but this is not observed here as cognitive tests such as the Y-maze or elevated plus maze did not evidence any significant improvement.

    Therefore, the understanding of the mechanistic link of pre-pathology CSF1R inhibition and pathology is somehow missing, and the extrapolation of these results to the preventative value of CSF1R blockade in AD must be with interpreted with caution.

    References:

    . Pharmacological targeting of CSF1R inhibits microglial proliferation and prevents the progression of Alzheimer's-like pathology. Brain. 2016 Mar;139(Pt 3):891-907. Epub 2016 Jan 8 PubMed.

    . Colony-stimulating factor 1 receptor inhibition prevents microglial plaque association and improves cognition in 3xTg-AD mice. J Neuroinflammation. 2015 Aug 1;12:139. PubMed.

    . Eliminating microglia in Alzheimer's mice prevents neuronal loss without modulating amyloid-β pathology. Brain. 2016 Apr;139(Pt 4):1265-81. Epub 2016 Feb 26 PubMed.

  2. The authors present an intriguing study that highlights the interplay between microglia and neurons in the progression of amyloid pathology in a mouse model of Alzheimer’s disease. Although others have demonstrated that early intervention with CSF1R inhibitors staves off neuronal loss, this study expands on the critical role microglia play in the progression of pathology.

    The authors report a remarkable reduction in plaque burden and intraneuronal Aβ after three months of PLX3397 treatment. One of the difficulties with the manuscript arises from the use of PLX3397 to ablate microglia, since it inhibits CSF1R/kit found both on microglia and in a subset of neurons. Thus, it is not possible to determine if the observed effects on pathology arise from salutary effects of the drug on neurons or from the loss of microglia in the brain. Surprisingly, no analysis on neuron number was done, which would be informative as to the impact of PLX3397 on neuronal health. This is an important aspect to understand, as it feeds into the authors’ proposal of plaques deriving from dying neurons. 

    Although PLX3397 effectively reduces the number of microglia and plaques in the brain, it is puzzling that the remaining plaques still have high microglia coverage, as the authors note in the discussion. These resilient microglia are engaged with the plaque surface, and raise the question of why they have become CSF1-independent for their survival.

    The study also implies that ablation-resistant microglia may be endowed with beneficial pathology-modifying effects that depletion therapies could harness. Microglial ablation strategies should be employed cautiously in a translational setting, since it is unknown if any detrimental side effects arise following long-term ablation. Indeed, mice and humans with loss-of-function CSF1R mutations have a broad range of pathological phenotypes.

    Overall this paper highlights some exciting ideas; however, it falls short on the delivery of clear results implicating microglia in the observed phenotypes and behavioral improvement.

  3. I am puzzled by this experimental observation of supposedly beneficial effects of interference with CSF1-R function given the clinical syndrome of HDLS/POLG, a dementing condition due to CSF1-R loss-of-function mutations (e.g., Pridans, et al. 2013). What may be possible explanations that could perhaps reconcile these two apparently opposite effects of CSF1-R?

    References:

    . CSF1R mutations in hereditary diffuse leukoencephalopathy with spheroids are loss of function. Sci Rep. 2013 Oct 22;3:3013. PubMed.

Make a Comment

To make a comment you must login or register.

References

Research Models Citations

  1. 5xFAD

News Citations

  1. Microglial Magic: Drug Wipes Them Out, New Set Appears
  2. In TDP-43 Mouse, Hesitant Microglia Eventually Swoop to the Rescue

Paper Citations

  1. . Eliminating microglia in Alzheimer's mice prevents neuronal loss without modulating amyloid-β pathology. Brain. 2016 Apr;139(Pt 4):1265-81. Epub 2016 Feb 26 PubMed.
  2. . Colony-stimulating factor 1 receptor (CSF1R) signaling in injured neurons facilitates protection and survival. J Exp Med. 2013 Jan 14;210(1):157-72. PubMed.
  3. . Depletion of microglia augments the dopaminergic neurotoxicity of MPTP. FASEB J. 2018 Jun;32(6):3336-3345. Epub 2018 Jan 22 PubMed.
  4. . A Systems-Level Framework For Drug Discovery Identifies Csf1R As A Novel Anti-Epileptic Drug Target. bioXriv, May 22, 2017

Further Reading

No Available Further Reading

Primary Papers

  1. . Early long-term administration of the CSF1R inhibitor PLX3397 ablates microglia and reduces accumulation of intraneuronal amyloid, neuritic plaque deposition and pre-fibrillar oligomers in 5XFAD mouse model of Alzheimer's disease. Mol Neurodegener. 2018 Mar 1;13(1):11. PubMed.